Regulatory Impact Analysis,
Oxides of Nitrogen Pollutant Specific Study
" " and •
Summary and Analysis of Comments
Control of Air Pollution from New Motor
Vehicles and New Motor Vehicle Engines:
Gaseous Emission Regulations for 1987
. and Later Model Year Light-Duty Vehicles,
and for 1988 and Later Model Year
Light-Duty Trucks and Heavy-Duty Engines;
Particulate Emission Regulations for 1988
and Later Model Year Heavy-Duty Diesel Engines
March 1985
Environmental Protection Agency
Office of Air and Radiation
Office of Mobile Sources
-------
Regulatory Impact Analysis,
Oxides of Nitrogen Pollutant Specific Study
and
Summary and Analysis of Comments
Control of Air Pollution from New Motor
Vehicles and New Motor Vehicle Engines:
Gaseous Emission Regulations for 1987
and Later Model Year Light-Duty Vehicles,
and for 1988 and Later Model Year
Light-Duty Trucks and Heavy-Duty Engines;
Particulate Emission Regulations for 1988
and Later Model Year Heavy-Duty Diesel Engines
March 1985
Environmental Protection Agency
Office of Air and Radiation
Office of Mobile Sources
-------
TABLE OF CONTENTS
Page
1. Introduction 1-1
I. Organization 1-1
II. Background of the Regulations 1-1
A. Clean Air Act Requirements 1-1
B. Regulatory History 1-2
III. Description of the Action 1-3
- -•• A. New Emissions Standards 1-3
B. Particulate and Nox Averaging - 1-4
- - - - - -— c. - New Allowable Maintenance Regulations . 1-4
D. Test Procedure Revisions 1-5
IV. List of Commenters - 1-5
2. Technological Feasibility 2-1
I. Introduction .2-1
II. Light-Duty Trucks (LDTs) 2-1
A. Synopsis of NPRM Analysis ". . 2-1
B. Summary and Analysis of Comments .... 2-3
C. Conclusions 2-20
III. Heavy Duty Gasoline Engines (HDGEs) 2-21
A. Synopsis of NPRM Analysis 2-21
B. Summary and Analysis of Comments .... 2-23
C. Conclusions 2-29
IV. Heavy Duty Diesel Engines (HDDEs) 2-30
A. Synopsis of NPRM Analysis 2-30
B. Summary and Analysis of Comments .... 2-34
C. Conclusions 2-70
s
-i-
-------
TABLE OF CONTENTS (confd)
Page
3. Economic Impact 3-1
I. Synopsis of NPRM Analysis . . . 3-1
A. Cost to Manufacturers 3-2
B. Cost to Users . . . .- . 3-3
II. Summary and Analysis of Comments 3-4
A. LOT NOx ' 3-4
B. HDGE NOx 3-20
C. HDDE NOx and Particulate Standards . . . 3-32
D. Socioeconomic Impacts • 3-99
4. NOx and Particulate Environmental Impact 4-1
I. Overview of NPRM Analyses 4-1
A. Oxides of Nitrogen (NOx) 4-1
B. Particulate Matter 4-2
II. Summary and Analysis of Comments on NPRM
Environmental Impact and Air Quality
Projections 4-5
A. Factors Common to Both Analyses .... 4—5
B. Factors Specific to NOx 4-14
C. Factors Specific to Diesel Particulate . 4-19
III. Emissions/Air Quality Projections 4-23
A. NOx Analysis 4-23
B. Diesel Particulate Analysis 4-37
5. Cost Effectiveness 5-1
I. Overview of NPRM Analysis 5-1
II. Summary and Analysis of Comments 5-2
III. Updated Cost Effectiveness Analysis 5-4
-ii-
-------
TABLE OF CONTENTS (confd)
Page
A. Changes in Analysis 5-4
B. Results of Updated Analysis 5-5
C. Comparison to Other Control Strategies . 5-11
D. Conclusion 5-16
6. Alternative Actions 6-1
I. .-Introduction 6-1
II. Alternative Light-Duty Truck (LOT) NOx
Standards 6-1
III. Alternative Heavy-Duty Engine (HDE) NOx
Standards 6-4
IV. Alternative Heavy-Duty Diesel Engine (HDDE)
Particulate Standards 6-6
Appendix A - Summary and Analysis of Comments
on the Proposed Particulate Test Procedure for
Heavy-Duty Diesel Engines A-l
I. Recommendations Accepted by EPA A-'l
II. Recommendations Not Accepted by EPA A-8
III. Issues Raised by EPA in NPRM A-21
Tt/ m-Vjojr Tcciiae ......... A —7.1
-111-
-------
CHAPTER 1
INTRODUCTION
I. Organization
As required by Executive Order 12291, this document has
been prepared to summarize the results of all analyses
conducted in support of the final rule for gaseous emission
regulations for 1988 and later model year light-duty vehicles,
light-duty trucks, and heavy-duty engines and for particulate
emission regulations for 1988 and later model year heavy-duty
diesel engines. In addition, this document also provides a
summary and analysis of most of the comments received in
response to the Notice of Proposed Rulemaking (49 FR 40258
October 15, 1984). Included here is a consideration of the
technological feasibility, economic impact, environmental
effects and cost effectiveness of the standards along with the
development of data on the impacts of several regulatory
alternatives. The remaining issues raised by commenters to
this rulemaking are reviewed and responded to in the preamble.
These include the proposed averaging program, allowable
maintenance provisions and high altitude standards. The oxides
of nitrogen (NOx) environmental impact analysis contained in
this document also serves as the NOx pollutant-specific study
required by Section 202(a)(3)(E) of the Clean Air Act.
The material presented in this document deals primarily
with those areas of the draft Regulatory Impact Analysis-[l3
which were the subject of public ^ comment. Areas of analysis
which were not commented upon are repeated here only where
needed to aid the understanding of material being revised. The
draft analysis is therefore incorporated into this document by
reference for treatment of topics not specifically re-addressed
herein.
II. Background of the Regulations
A. Clean Air Act Requirements -
The Clean Air Act Amendments of 1977 created a statutory
heavy-duty vehicle (HDV) class and established mandatory
emissions reductions for that class. Under the language of the
amendments, all vehicles over 6,000 Ibs gross vehicle weight
(GVW) were defined as "heavy duty" and were required to achieve
a 75 percent reduction in NOx emissions from uncontrolled
levels, effective with the 1985 model year.
The Act made no specific provisions for light-duty trucks
(LDTs), which at that time only encompassed LDTs between 0 and
6,000 Ibs GVW (light LDTs). These LDTs were regulated by EPA
as a separate class under the general authority of the Clean
Air Act. Beginning with the 1979 model year, EPA expanded its
standards for the LDT class to 8,500 Ibs GVW, thus encompassing
those heavy LDTs (6,001 to 8,500 Ibs GVW) which are subject to
the heavy-duty vehicle provisions mentioned above.
-------
1-2
The Act also authorizes the Administrator to temporarily
establish revised NOx standards for heavy-duty engines if the
statutory standards cannot be achieved without increasing cost
or decreasing fuel economy to an "excessive and unreasonable
degree."T2] The new heavy-duty engine NOx standards in this
document are being promulgated under these provisions of the
Act.
The amendments of 1977 also require the "greatest degree
of Cparticulate] emissions reduction achievable," given the
availability of control technology and considering cost,
leadtime and energy impacts.[3] These reductions were to begin
in the 1981 model year. Although not specifically limited as
to applicability in the language of the Amendments of 1977, it
was recognized that the requirement was aimed at diesel
engines. The heavy-duty diesel engine (HDDE) particulate
standards in this rulemaking are based on this authority.
B. Regulatory History
The first NOx standards antedated the amendments of 1977*.
Prior to the 1975 model year, LDTs complied with the 3.0 g/mi
NOx standard that had been established two years earlier for
LDVs. With the splitting off of the LDT class for the 1975
model year, LDTs were required to meet a NOx standard of 3.1-
g/mi, comparable in stringency to the LDV standard. Heavy-duty
engines (HDEs) had no separate NOx standard until the 1985
model year, however, there have been combined hydrocarbon (HC)
+ NOx standards in place for HDEs since the 1974 model year.
The current NOx standard for 1979 and later model year
LDTs is 2.3 g/mi, comparable in stringency to the 2.0 g/mi
standard established for LDVs of that year. Beginning in 1979,
the LDT class was expanded to include vehicles between 6,001
and 8,500 Ibs GVW. The current NOx standard for HDEs is 10.7
g/BHP-hr, established originally for the 1984 model year, but
later made optional until the 1985 model year.
Turning now to more recent actions, an Advanced Notice of
Proposed Rulemaking (ANPIW) was promulgated for LDT and HDE NOx
emissions in January of 1981 (46 FR 5838). Standards of 1..2
grams per mile for LDTs and 4.0 g/BHP-hr for HDEs were
suggested effective for the 1985 and 1986 model years,
respectively. These standards did not correspond to the
statutory 75 percent reduction as noted above, but were
proposed because they were comparable in stringency to the
existing 1.0 g/mi LDV NOx standard in the case of LDTs and
because they represented what EPA believed at that time to be
the lowest practicable standard given the available technology
in the case of HDEs.
-------
1-3
The first diesel particulate standards were established
for LDVs and LDTs, effective beginning with the 1981 model
year. A standard of 0.60 g/mi was established for both LDVs
and LDTs, representing an achievable level for the (then)
available technology. More stringent standards (at 0.26 for
LDTs, and 0.20 for LDVs) were also promulgated effective
beginning with the 1985 model year, but these have been delayed
and will now be effective for the 1987 model year (49 FR 3010,
January 24, 1984). For HDDEs, a Notice of Proposed Rulemaking
(NPIW) was published in January, 1981 (46 FR 1910) which-
proposed a standard of 0.25 g/BHP-hr for 1986 and later model
years.
Because of the related technical issues that were raised
during the comment periods for both the NOx ANPRM and the
-particulate NPR4 and the interrelationship between NOx and
particulate emissions, EPA decided to issue a combined NPRM to
address these issues to insure that manufacturers could direct
their efforts at meeting a unified set of emission standards.
The Notice of Proposed Ruleraaking was published on October 15,
1984 (49 FR 40258). This final rule, preceeded by public
hearings and a public comment period, completes the rulemaking
process.
Ill. Description of the Action
A. New Emissions Standards - ' ,
This rulemaking contains new low-altitude NOx standards
for LDTs and HDEs, new low-altitude particulate standards for
HDDEs and new high-altitude idle Co, NOx and particulars
standards for LDTs. For 1988 and later model years, the NOx
standard for LDTs is 1.2 g/mi for LDTs up to and including
3,750 Ibs loaded-vehicle weight. The standard for 1988 and
later model year LDTs over the above weight limit is 1.7 g/mi.
A staged NOx standard is established for HDEs to allow leadtirae
for further development of control technology. The NOx
standard for 1988-90 model year HDEs is 6.0 g/BHP-hr,
representing a level that is achievable given the available
leadtime for engines currently in production, with a more
stringent standard of 5.0 g/BHP-hr effective for 1991 and later
model year engines.
A three-phased particulate standard is established for
HDDEs. Model year 1988-90 HDDEs will meet a standard of 0.60
g/BHP-hr. For 1991-93 model years, urban bus engines will
comply with a standard of 0.10 g/BHP-hr, while the remaining
HDDEs will meet a standard of 0.25 g/BHP-hr. Both of these
1991 standards will likely require the use of trap oxidizers on
a majority of applications. This will be followed by the third
phase, when all 1994 and later model year HDDEs will comply
with the 0.10 g/BHP-hr standard.
-------
1-4
Finally, certain new high altitude standards are
established for light-duty trucks. NOx standards equal to the
1.2 g/mi and 1.7 g/mi low-altitude standards are established,
along with an idle CO standard of 0.50 percent of exhaust gas
flow at idle (gasoline-fueled light-duty trucks only) and a
particulate standard of 0.26 g/rai (diesel light-duty trucks
only).
B. Particulate and NOx Averaging
With this rulemaking, particulate averaging will be
afforded to manufacturers of 1991 and later model year HDDEs.
However, they will not be allowed to average HDDEs with LDDTs
or LDDVs if the manufacturer' s product line also includes these
vehicle types. Similarly, averaging California engines with
engines intended for sale in non-California areas will not be
permitted, although averaging within each of these areas is
allowed. Urban buses will be excluded from the particulate
averaging program to insure the maximum reduction in urban
particulate emissions. Because HDDE standards are expressed in
mass per unit of work (g/BHP-hr) .rather than mass per unit of
distance travelled (g/mi) and because HDDEs are divided into
subclasses with widely varying useful life periods, averaging
will be limited to within each of the existing subclasses
(light-, medium-, and heavy-heavy duty) and the calculation of •
average particulate emissions must include weighting factors
for brake horsepower as well as for production volume.
NOx averaging has been established for 1991 and later
model year HDEs and is similar to the particulate averaging
program, with the following exceptions. The NOx averaging
program is restricted by fuel type, with gasoline-fueled and
diesel engines complying with the standard by separate
averages. For HDDEs, the averaging is restricted by engine
subclass (light, medium, and heavy); however, gasoline-fueled
HDEs have no such restriction. Also, urban buses excluded from
particulate averaging may be included in the NOx averaging
program for all HDEs.
Finally, NOx averaging for light-duty trucks is
established beginning in 1988. This program is patterned
closely after those established for heavy-duty engines and the
existing light-duty diesel averaging program. Further details
for both the NOx and particulate averaging programs are
outlined in the preamble and included in the revised
regulations.
C. New Allowable Maintenance Regulations
The allowable maintenance provisions proposed have been
retained largely unchanged. The concept of emission- and
non-emission-related maintenance has been extended from LDTs
-------
1-5
and HDEs to encompass LDVs as well. Maintenance intervals have
been changed, including revisions to the proposed intervals, as
outlined in the preamble. Manufacturers will be required to
demonstrate the likelihood of in-use performance for certain
critical emission-related maintenance.
D. Test Procedure Revisions
The heavy-duty engine test procedures have been revised to
incorporate particulate test procedures. These include changes
in response to comments along with other minor corrections, as
outlined in the preamble to the final rule.
IV. List of Commenters
The following individuals, organizations, public
authorities and manufacturers submitted written comment in
response to the NPRM (49 FR 40258). This list contains only
those comments received by January 4, 1985. Comments received
after that date, although not specifically identified here,
have also been incorporated fully into EPA1s analyses along
with those listed.
1. Adair, Holiday, Akron, OH
2. American Automobile Association
3. American Honda Motor Company
. 4. American Lung Association
5. American Lung Association, of Berks County (PA)
6. American Lung Association of Deleware/Chester
Counties (PA)
7. American Lung Association of Florida
8. American Lung Association of New Jersey
9. American Lung Association of Western Missouri
10. American Motors Corporation (AMC)
11. American Public Transit Association
12. Arent, Fox, Kinter, Plotkin & Kahn for MEMA
13. Arizona Lung Assoc.
14. Audubon Society of Ohio
15. Automobile Importers of America
16. Bass, Jean, Ross, CA
17. Baughman, Jon, Bedford Hgts., OH
18. Baumgarten, Sara, Bridgewater, MA
19. Bergen County (NJ) Audubon Society
20. Bickford, Isabel, Williamsville, NY
21. Biesterfeld, Cathy, Homewood, IL
22. Bradman, Asa, Ross, CA
23. Brenner, Jeff, New Brunswick, NJ
24. Brown, Bruce and Sharon, Chicago, IL
25. Brown, Paul M., Sun Olympiad '80
26. Burchard, Ann, Robert and Rachel, Catonsville, MD
27. California Air Resources Board
28. California Dept of Justice
-------
1-6
29. Callan, Ida, Vienna, OH
30. Cape Henry Audubon Society
31. Capital District Transportation Authority
32. Caterpillar Tractor Company
33. Chemel, Bonnie, Evans City, PA
34. Chicago Transit Authority
35. Chrysler Corporation
36. Ciak, Josephine, North Arlington, NJ
37. Clark County (NV) Health District
38. Coalition for Clean Air
39. Coalition for the Environment
40. Colorado Department of Health
41. Connaughton, Ruth K.
42. Cummins Engine Company
43. Delaware Valley Citizens Council for Clean Air
44. Delello, Michael, Saranac Lake, NY
45. Dillon, Mary, Elma, NY
46. Dolinka, Marvin & Toby, Grand Rapids, MI
47. East Michigan Environmental Action Council
48. El Paso Clean Air Coalition
49. Environmental Alternatives, Inc.
50. Faulconer, Mrs. James H., Strasburg, VA
51. Fisher, C. Donald, Muncy, PA
52. Ford Motor Company
53. Fox, Warren, Linwood, NJ
54. Gardiner, Jeffrey, Schenectady, NY
55. General Motors Corporation (GM)
56. Geymer, Christine, Oak Park, IL
57. Gordon, Robin, Great Neck, NY
58. Greater Cleveland Regional Transit Authority
59. Grenfo, Louise, Crossville, TN
60. Group Against Smog and Pollution
6L. Hamilton, James, Cleveland, OH
62. Hawarth, Terrie, E. Grand Rapids, MI
63. Holmes, David, Clarion, PA
64. Humphreys, Betsy, Morgantown, WV
65. Huser, Bill, So. Sioux City, NE
'66. International Harvester Company (IHC)
67. Isker, C., Buffalo, NY
68. Iwanik,_Mike, Richmond, VA
69. Jaguar Cars Inc.
70. Jenner & Block for Engine Manufacturers Association
(EMA)
71. Joan Katz Productions
72. Johnson, David, Pueblo, CA
73. Johnson, Nina, Boulder, CO
74. Johnson, Rose Mary, Louisville, KY
75. Kemp, Katherine, Chicago Heights, IL
76. Kulakowski, Lois, Tucson, AZ
77. LTV Aerospace and Defense Company
78. League of Women Voters of the Clemson Area
-------
1-7
79. League of Women Voters of the Pittsburgh Area
80. Lewis, Nana, Larkspur, CA
81. Love, John, Boulder, CO
82. Mack Trucks Inc.
83. Mannchen, Brandt, Houston, TX
84. Mansell, Gerda, Lancaster, NY
85. Manufacturers of Emission Controls Assoc.
86. Massachusetts Department of Env. Quality Engineering
87. Mazda (North America)
88. McCarty, Donna, Indianapolis, IN
89. McGuire Clinic
90. Mercedes-Benz Truck Co.
91. Meyer, Arthur, Akron, PA
92. Motor and Equipment Manufacturers Association
93. Motor Vehicles Manufacturers Association (MVMA)
94. Mueller, Catherine & Edwin, Buffalo, NY
95. Mueting, Ann, Plymouth, MN
96. Murkeloff, Robert, Houston, TX
97. NJ Transit Bus Operations
98. Natural Resources Defense Counsel (NRDC)
99. New Jersey Department of Environmental Protection
100. New Mexico Environmental Improvement Division
101. Newberry, William, Grand Rapids, MI
102. Nissan Research & Development
103. Oakes, Margaret, Boulder, CO
104. Oregon Department of Environmental Quality
105. Osterpard, Elsie, Grand Rapids, MI
106. Otter Creek Audubon Chapter
107. PACCAR Inc.
108. Pettit, Marie, Harrisonburg/ VA
109. Rhode Island Department of Environmental Management
110. Richmond (VA) Audubon Society
111. Rolls-Royce Motors
112. Rosche, Olga, South Wales,NY
113, Ross, G.M., Lowell, MI
114. STAPPA/ALAPCO
115. Saab-Scania of America
'116. Schiffirth, Anne & Jim, Pittsburgh, PA
117. Schoenfeld, Josephine, Grand Island
118. Sherman, L. Ann, Schaumburg, IL
119. Shutter, S.L.
120. Simpson, Robert, Flint, MI
121. Smith, Bertha, Grand Rapids, MI
122. South Coast Air Quality Management District
123. Southern California Rapid Transit District
124. St. Cloud Area Environmental Council
125. Storgul, Pauline, Chicago. IL
126. Tonseth, Phebe, Cumberland Foreside, ME
127. Toyota Technical Center, USA
128. U.S. Department of Energy
129. U.S. DOT, Urban Mass Transit Administration
-------
1-8
130. VIA Metropolitan Transit
131. Volkswagen of America
132. Volvo-North American Car Operations
133. Volvo White Truck Corporation
134. Volvo of America Corporation - Bus Division
135. Washington State Department of Transportation
136. Wedow, Nancy, Palatine, IL
137. West Michigan Environmental Action Council
138. White Lung Association
139. Willard, Dwight, Albany, CA
140. Williams, Mark, San Francisco, CA
-------
1-9
References
1. "Draft Regulatory Impact. Analysis and Oxides of
Nitrogen Pollutant Specific Study Control of Air Pollution From
New Motor Vehicles and New Motor Vehicle Engines: Gaseous
Emission Regulations for 1987 and Later Model Year Light-Duty
Vehicles, Light-Duty Trucks, and -Heavy-Duty Engines;
Particulate Emission Regulations for 1987 and Later Model Year
Heavy-Duty Diesel Engines," U.S. EPA, OAR, OMS, October 1984.
2. The Clean Air Act As Amended August 1977, Serial No.
95-11, Section 202(a)(3)(B)&(C).
3. IBID; Section 202 (a)(3)(iii); p. 102. ,
-------
CHAPTER 2
TECHNOLOGICAL FEASIBILITY
I. Introduction
This chapter analyzes the technical feasibility and the
leadtime requirements of the final 1988 and later model year
light-duty truck (LOT) standards for oxides of nitrogen (NOx)
emissions and the final heavy-duty engine
-------
2-2 .
It was concluded from this analysis that the 1.2 g/mi NOx
standard was feasible for all LDGTs and for lighter LDDTs
(LDDTiS). The principal compliance means for LDGTs would be
closed loop, three way catalyst technology, while lighter LDDTs
would rely on the application of EGR. For LDDT2s it was
concluded (because of their heavier weights and larger frontal
areas) that a 1.2 g/mi NOx standard would increase particulate
levels such that it would affect the stringency of the 0.26
g/mi particulate standard. Consequently, EPA then considered a
1.7 g/mi NOx standard for LDDT2s.- As in the case of
LDDTiS, EGR was projected to provide the means for compliance.
It was concluded that a 1.7 g/mi standard appropriately
balanced the need for NOx and particulate control. As a
result, EPA decided to propose a. NOx standard of 1.7 g/mi for
LDDT:S. EPA also decided that it was most equitable to
propose the 1.7 g/mi NOx standard for LDGT2s. This was done
because a more stringent NOx standard for heavier LDGTs would
encourage the purchase of LDDT2s, which EPA did not wish to
do. Further, the loss of NOx control due to a more lenient
standard for LDGTzS was small compared to the case where only
diesels were affected.
Considerations of the effects on fuel economy of the
proposed NOx standards in combination with the technologies
expected to be employed led to the conclusion that LDGTs which
were converted to three-way catalyst technology from oxidation
catalyst technology could experience up to an 8 percent
improvement in fuel economy. For those LDGTs which already
employed three-way catalyst technology, it was projected that
some small fuel economy loss might occur. It was forecasted on
a sales-weighted basis that roughly a 2-4 percent improvement
in fuel economy would be associated with the proposed standards
for LDGTs. For LDDTs, consideration of the effects of the
proposed standard on fuel economy led to the conclusion that no
significant fuel economy penalty would result from the proposed
standards for either LDDT,s or LDDTzs. This conclusion was
based on evaluations of the differences in fuel economy between
LDDT,s with and without EGR and on the benefits associated
with the use of electronic controls on LDDT2s.
For LDTs sold at high altitude, EPA proposed the same NOx
standards' as were proposed for low altitude LDTs because NOx
emissions do not tend to increase with altitude. An idle' CO
standard for high altitude LDTs equal to the 0.50 percent
standard already required for low altitude LDTs was proposed
because a 90 percent reduction from baseline high altitude idle
CO levels resulted in a numerical value of 0.51 which, when
rounded, was equal to the low altitude value. A particulate
standard equal to that at low altitude was also proposed.
-------
2-3
B. Summary and Analysis of Comments
1 . Introduction
The comments received concerning the feasibility of the
proposed LOT NOx standards are summarized and analyzed below.
When more than one commenter raised the same basic issue, the
issue is treated once in the summary and analysis with
identification of the multiple sources of the comment. While
the proposed standards are applicable to two types of engines
(gasoline and diesel) used to power LDTs and are numerically
different as a function of the weight of the LOT, comments will
be treated by issue with appropriate consideration of these
distinctions where necessary. The issues contained in the
summary and analysis of comments which follows are, the
technical feasibility of the proposed standards, the leadtime
required for compliance, the effect of the standards on fuel
economy and other minor issues.
2 . Technical Feasibility of the Proposed Standards
Six commenters provided comments on the technical
feasibility of the proposed standards with respect to
gasoline-fueled LDTs. Chrysler, Ford, General Motors, Nissan
and Toyota stated that the proposed standards (1.2 g/mi for
LDGTiS and 1.7 g/mi for LDGT2s> were technologically
feasible. ' • - - --•-
VW disagreed with the technical feasibility of the
proposed standard on the grounds that the allowable maintenance
provisions and the full-life useful life requirement result in
requirements which are beyond the capabilities of current
in-use or reasonably forseeable technology. VW did not,
however, provide any information in substantiation of thsir
statement. Lacking substantiating information for the VW
statement and considering the position taken by the other
commenters leads to the conclusion that the proposed standards
are technologically feasible for LDGTs using the technologies
identified in the Draft RIA (three way catalyst and closed loop
fuel control). This is confirmed by certification data for the
1985 model year. As shown in Tables 2-1 to 2-7, nearly all
LDTs certified with three-way closed loop technology are
already in compliance with 1.2/1.7 standards. Both full-life
useful life and revised allowable maintenance provisions apply
to federally certified LDTs for 1985.
In the case of diesel fueled LDTs, three
provided comments on the technological feasibility of -.;
proposed standards (1.2 g/mi and 1.7 g/mi). One cornmer.ter
Ford, stated that the proposed NOx standards ~et
-------
2-4
•Bible
1985 49-State Federal Certification Data
for LDGTs Equipped with Three-Way Catalyst Technology
Manufacturer
Light-Duty Trucks
American Motors
Pord
Nissan
Mitsubishi
Suzuki
Volkswagen
Engine Family
- Class 1
F7M2.5TIHEA5
FP42.3T5FAG7
FN52.4T9FAFO
IMT2.0T2FFDX
P4T2.6T2FFD2
FSK1.0T1FSF7
FVW1.9T5CVF8
Emission
Control
Technology*
EGR/PLS/3CL
E3H/3CL
- B3H/3CL/OTR
B3R/PLS/3CL/OTR
EER/PLS/3CL/OTR
3CL
3CL
Mean
NDx DF
1.541
1.057
1.095
1.043
1.100
1.216
1.043
Mean Cert.
NOx Level
1.27
0.34
1.11
1.12
1.38
. 1.70
0.74
Light-Duty Trucks - Class 2
American Motors
Ford - -
F7M4.2T2HEA7
FA45.9T2HLE2
FIM5.0T5HAG8
FM5.8T2H3G1
FW5.8T4GAF4
BSR/PLS/OXD/3CL
BGR/EMP/3WY
EX3R/WP/OXD/3CL/OTR
BGR/P4P/CKD/3CL
ECR/WP/OXD/3WY
1.114
1.056
1.021
1.005
1.131
1.4
1.35
0.80
1.85
2.1
E3R = Exhaust gas recirculation.
3CL = Three-way catalyst, closed-loop fuel control,
P4P = Air punp.
OTR = Other.
PLS = Pulse air injection.
3WY = Three-way catalyst, open-loop fuel control.
OXD = Oxidation catalyst.
El = Engine modification.
-------
2-5
Table 2-2
1985 49-State Federal Certification Data
for LDGTs Equipped Without Three-Way Catalyst Technology
Manufacturer
Light-Duty Trucks
American Motors
Chrysler
Ford
^General Motors
Isuzu
Toyota
Fuji
"•
Light-Duty Trucks
Chrysler
General Motors
Engine Family
- Class L
FAM2.8T2AXE3
FCR2.2T2AAB6
FCR2.6T2AAB8
FFM2.0T1AGF2
FIG1.9T2HJC2
FIG2.8T2HTX1
FSZ119T2AAG2
FTY2.4T2AFF1
FFJ1.8T2AFJ1
FFJ1.8T2AFK2
- Class 2
FCR3.7T1BBAO
FCR5.2T2BBF6
FCR5.9T4B3F1
FIG4.3T4HHC1
FIG5.7T4HHCO
Emission
Control
Technology*
EGR/PMP/OXD
EGR/PMP/OXD
EGR/PMP/OXD
EGR/PMP/OXD
EGR/PMP/OXD
EGR/PMP/OXD
EGR/P4P/OXD
EGR/PLS/OXD/OTR
EGR/PLS/OXD
EGR/PLS/OXD
EGR/PMP/OXD
EGR/PMP/OXD
EGR/PMP/OXD
EGR/PMP/OXD
EGR/PMP/OXD
Mean
NOx DF
1.050
•i.o -
1.0
.1.0
1.0
1.050
0.975
1.10
1.017
1.017
1.0
1.0
1.0
1.012
1.0
Mean Cert.
NOx Level
1.7
1.57
1.2
1.95
1.9
1.60
1.70
1.85
1.25
1.75
1.9
1.9
1.62
1.45
1.73
Toyota
FTY4.2T2AFF5
EGR/P1P/OXD/OTR
0.816
0.99
See Table 2-1 for definition of terms.
-------
2-6
Table 2-
49-State Federal Certification Data for L985 LDDTs
Manufacturer
Light-Duty Trucks
American Motors
Ford
General Motors
Grumman Olson
Isuzu
Nissan
Mitsubishi
Toyota
Light-Duty Trucks
Engine Family
- Class 1
FSM2.1K6JZT7
FFM2.3KJAF1
FIG2.2K7ZZ98
FGR1.6K6JAA6
FSZ137K6JCD5
FNS2.5K6JAF7
EMT2.3K6JFD2
FTY2.4K6JFF1
FTY2 . 4K6 JFT9
- Class 2
Emission
Control
System*
EM
EM
EM
EM
EM
EM
EM
EM
EM
NOx DF
1.038
0.995
1.000
1.000
0.956
1.015
0.995
1.005
1.000
Mean Cert
NOx Level
1.40
1.65-
1.80
1.10
1.70
1.80
1.85
1.70
1.80
General Motors FIG6.2K7ZZ42
EGR
1.013
1.90
See Table 2-1 for definition of terms.
-------
2-7
•Sable 2-4
1985 California Only Certification Data
. for LDGTs Equipped with Three-V&y Catalyst Technology
Emission Control
Manufacturer
Bngine Family
Light-Duty Gasoline Trucks - Class
Fuji
Toyota
Ford
Isuzu
MC
Nissan
General Motors
Volkswagen
FFJ1.8T2HCGO
FFJ1.8T2AFK2
FFJ1.8T2HCPO
FTY2.4T2FCC7
FPM2.3T5FFG6
FEM2.8T2HKGO
FSZ119T2FDG1
FAU50TIHEA6
FA1173T2F4C3
FNS2.4T9FAC8
FIG2.ST2TPA3
FVW1.9T5CVC5
Light-Duty Gasoline Trucks - Class
»IC
Ford
Chrysler
General Motors
piM 2 5ST2HEAO
FA1360T2HLEO
FP14.9TIHGG6
FF45.8T2B3G1
FCR3.7TIHDS1
FIG4.3T4TAA3
FIG5.7T4TYS6
Technology*
1
B3R/PLS/OXD/3CL
E3R/PLS/OXD/3CL
B3R/PLS/OXD/3CL
EGP/3CL/OTR
EGR/3CL
B3R/P1P/OXD/3CL
B3R/EMP/3CL
B3R/PLS/3CL
B3R/EMP/3CL
BGP/3CL/OTR
S3R/P-IP/3CL
3CL
2
BSR/HS/3CL
EGR/WP/3WY
B3R/EMP/OXD/3CL
EGF/EMP/OXD/3CL
33R/OXD/3CL
EGR/WP/3CL
EGR/P-IP/3CL
NOx DF
1.293
1.293
1.293
1.076
1.000
1.091
.997
1.372
.892
1.159
1.000
.526
1,045
1.050
1.084
1.100
.947
1.031
1.031
Mean Cert.
NOx Level
0.33
0.31
0.28
0.44
0.27
0.70
0.43
0.83
0.94
0.71
0.64
0.63
0.68
0.56
0.73
0.81
0.82
0.32
0.60
fl-lG
FAZ4.2T2HFGO
BGR/OXD/3CL
1.170
0.53
See Table 2-1 for definition of terms.
-------
2-8
•Bible 2-5
1985 California Cnly Certification Data for LDDTs
M anu factur er
Emission Control
Engine Family Technology*
NOx DP
Mean Cert.
NOx Level
Light-Duty Diesel Trucks - Class 1
Isuzu
Mitsubishi
Toyota
Nissan - - -
General Motors
Light-Duty Diesel
General Motors
General Motors
FSZ137K6JBD3
IMT2.3K6JCB5
PTY2.4K6JCT3 OTR
FNS2.5K6JAC4
FIG2.2K7ZZL2
Trucks - Class 2
FIG6.2K7ZZ75
FIG6.2K7ZZ75
1.034
.964^
1.043
.939
1.000
1.015
1.015
.84
.76
.86
.75
.89
1.30
1.70
See Table 2-1 for definition of terms.
-------
2-9
•Bible 2-6
L985 50-State Certification Data for LDGTs
Equipoed wich 3-Vfay Catalyst Technology
Manufacturer Engine Family
Light-Duty
RAG
Chrysler
CM
Mitsubishi
Toyota
Light-Duty
Button
' Ford
Winnebago
• Zi inner
CM
Chrysler
Trucks - Class 1
FAZ2.5T1HRG9
-~ PCR2.2T2HB49
FCR5.2T2KBN1
F2G2.575TPG9
EMT2.0T2FCA1
PAT2.6T2FCM
FTJ2.4T5FBT6
FTY2.0T5FBB3
FT52.4T5FBB5
Trucks - Class 2
„ FDN4.LT5NKA4
. , i FIMS.OTSHMS
FfoB2 . 2T3FGAO
, F^12.6T6FXX5
FZG2.5T5TFG9
"" FCR5.2T2HB^a
Emission
Control
Technology*
H3R/PLS/OXD/3CL
E3R/EMP/OXD/3CL
• B3R/EMP/OXD/3CL
BGR/3CL
EGR/PLS/3CL/OTR
EGR/PLS/3CL/OTR
EGP/3CL/OTR
EGR/3CL/OTR
EER/3CL/OTR
B3R/EMP/OXD/3CL
BGR/W P/OXD/3CL/OTR
E3R/3CL
3CL/OTR
B3R/3CL
E3P/WP.OXD/3CL
Mean
NOx DF
1.17
1.087
k
1.320
1.028
1.072
.683
1.568
1.463
1.157
1.100
1.170
1.150
1.320
1.051
1.158
Mean
Cert.
NOx Level
.38
.72
.42
.58
.68
.07
.17
.17
.48
.57
.61
.84
.23
.SI
.60
See Table 2-1 for definition of terms.
-------
2-10
•Bible 2-7
1985 50-State Certification Data for LDDTs
Manufacturer
Engine Family
Emission
Controls
Mean
NOx DF
Mean
Cert.
NOx Level
Light-Duty Trucks - Class 1
Mitsubishi
Nissan
Isuzu
Toyota
04
AdC
Ford
Grumman Olson
FdT2.3K6JCB5
IMT2.3K6JFD2
FSS2.5K6JFD2
FNS2.5K6JAF7
FSZ137K6JAF7
FSZ137K6JCD5
FTQ.4K6JCT3
FTX2.4K6JFF1
FTT2.4K6JFT9
FIG2.2K7ZZL2
FIG2.2K7ZZ98
FJM2.LK6JZT7
FEM2.3K6JAF1
FGR1.6K6JAA6
HER
EA
EGR
1M
B3R
Rd
BGR/OTR
El
H<
E3R
M
El
Bd
Bd * '"
.964
.995
.939
1.015
1.034
.956
1.043
1.005
1.000
1.000
1.000
1.038
.995
1.000
.75
1.85
.75
1.80
.84
1.70
.86
1.70
1.80
.89
1.80
1.40
1.50
1.10
Light-Duty Trucks - Class 2
Ford
CM
i
FW2.3K6JAF1
FIG6.2K7ZZ75
FIG6.2K7ZZ42
Ed
EGR
S3R
.995
1.015
1.013
1.80
1.50
1.90
See Table 2-1 for definition of terms.
-------
2-11
technologically feasible for LDDTs. Nissan stated that the
standard proposed for LDDT.s ( L. 2 g/rr.i) was feasible for some
of its two-wheel drive vehicles, while its other LDDT,s would
need new EGR systems to control NOx while simultaneously
complying with the particulate standard of 0.26 g/mi. CM
stated that the ability of the 6.2 liter engine to comply with
a NOx standard of 1.7 g/mi through the use of electronically
controlled EGR would be marginal and would result in increased
particulate emissions and a fuel economy penalty. GM provided
a figure (Figure III-C-1 in the GM comments) depicting the NOx
vs. particulate engine-out emission characteristics for the 6.2
liter engine in support of its position that particulate
emissions would greatly increase under a 1.7 g/mi standard and
jeopardize its ability to meet the LOT particulate standard of
0.26 g/mi.
Since all commenters concurred either fully or with some
qualification with the technological achieveability of the
proposed standards for LDDTs, analysis of the comments need
focus only on the qualifying statements presented by the
commenters.
In the Draft RIA (page 2-22), EPA concluded that all
Federal LDDT,s could be brought into compliance with the
proposed NOx standard (1.2 g/mi) through the use of EGR system
designs already being used on counterpart LDDTis certified to
the California NOx standard. The thrust of the Nissan comment
is that a new EGR system would have to be designed for use on
some of its LDDTts rather than the transfer of an existing
EGR system. Nissan is, therefore, concurring with the
technological feasibility of the proposed standard by the use
of EGR but is identifying a need for leadtime to design a new
EGR system (leadtime requirements are addressed later).
In responding to the comment from GM, EPA has plotted 1S85
model year certification data for the Federal and California
versions of the 6.2 liter GM engine on the curve provided by GM
in its Figure III-C-1 (reproduced here as Figure 2-1). These
values are shown as F,, and F2 (1.90 g/mi NOx, 0.41 g/ni
particulate and 1.90 g/mi NOx, 0.34 g/mi particulate) for the
two federal vehicles using EGR and as Ci and d (1.7 g/rni
NOx, 0.32 g/mi particulate and 1.3 g/mi NOx, 0.32 g/'nu
particulate) for the two California engines using
electronically controlled EGR. As can be seen from Figure 2-1,
NOx and particulate emissions actually being, achieved are
substantially lower than the gen-eralized curve presented by
GM. In addition, because the 6.2 liter GM engine is already
very close to the particulate standard on an engine-out basis,
EPA can see no basis for GM's concern about meeting tr.e
particulate standard. The application of particulate traps co
-------
Figure 2-1
NOx - Particulate Trade - Off Curve —- 6.2L Light Duty Diesel
I)
O,
1
0.80--
0.70 --
0.60 - -
0.50
0.40 •-
r.2
r.e
Electronic Modulated EGR
2.0
.2
2:4
Ni)x — Grams Per Mile
-------
2-13
approximately percent of these engines would, with averaging,
secure compliance with the 0.26 g/mi standard. EPA concludes
t'hat the NOx/particulate curve supplied by GM is not applicable
to current versions of its 6.2 liter engine, and that GM should
have no difficulty meeting a 1.7 g/mi standard with this engine.
3. Leadtime Requirements
Four commenters stated that the time required for
implementation of the LOT NOx standards for gasoline-fueled
vehicles exceeds the time available under the proposal.
Specifically, the comments were as follow. American Motors
"indicated that 34 months are normally required for the change
in catalyst technology required by the proposed standards.
General Motors stated that 106 weeks were required to make
changes to LOT bodies necessary to accommodate the larger
catalysts required for compliance with the proposed standards.
Nissan indicated that 2-1/2 to 3 years would be required for
changes to the vehicle body necessitated by the use of larger
catalysts. Toyota stated that -additional time is needed beyond
that available for compliance by the 1987 model year because of
the change in catalyst technology necessary and the need to
establish durability and reliability characteristics of the new
catalyst.
Two of the commenters (GM and Nissan) predicated their
leadtime requirements on the need to change the floorpan of the
vehicle body to accommodate larger catalysts. The tasks
required for the specified change in the vehicle body would be
the redesign of the floorpan to provide the necessary space and
the procurement of new dies for the manufacture of the
redesigned floorpans. Since the commenters indicated in other
areas of their comments that they have already quantified the
catalyst volume requirements, redesign of the fioorpan can oe
assumed to start essentially at the publication date of the
final rule. The maximum time requirement which could be
allocated for the redesign of the floorpan is 6 to 9 months.
Following redesign of the floorpan, the time required for the
procurement of new dies is between 26 and 36 weeks (Reference 2
at page 7-7) including the time required for installation of
the new dies in the presses. In total, these two tasks are
expected to require between 50 and 72 weeks or a maximum of 18
months. Starting with a publication date of March 15, 1985 for
the final rule and ending with October 1, 1986 for the
introduction date of the 1987 model year LDTs, defines a period
of 18-1/2 months for the execution of the tasks necessary cor
vehicle body redesign. Leadtime requirements for the redesign
of LDT bodies is, therefore, not a viable basis for clai-ir.g
that insufficient time is available for the implementation ;f
the proposed standards on LDGTs.
-------
2-14
Turning to the comments provided by American Motors, that
34 months is the normal requirement for the introduction of new
catalyst technology to a specified class of vehicle, the steps
required for this work can be identified as follows. The first
step would be the development of an overall system design
including the integration of the new system into the vehicles.
The subsequent steps would be: 1) ordering of modified tooling
for the manufacture of redesigned components which could
include vehicle body redesign, 2) the construction and testing
of experimental systems at several calibrations to establish
one or more calibrations for use on emission durability
vehicles, 3) collection of emission durability data, 4)
collection of data from emission data vehicles, and 5) -the
incorporation of modified tooling on machine lines for the
manufacturer of redesigned components.
EPA's timing requirement estimates for the primary tasks
in the critical timing path for gasoline-fueled LDTs are as
follows:
Task Time Requirement
Overall System Design and Vehicle Integration 4-6 months
Develop Durability Vehicle Calibrations 5-7 months
Generate Emission Durability Data 11-12 months
Develop Final Calibrations 1-2 months
Run Data Vehicles 1 month
Complete Certification Process
With EPA and Add Modified Tooling 1-2 months
TOTAL 23-30 months
Starting with a publication date of March 15, 1985 for the
final rule establishing the NOx standard, approximately 18-1/2
months is available prior to the start of the 1987 model year
in October of 1986. Since the minimum time required to perform
the necessary tasks is greater than the time available, it is
concluded that insufficient time was allowed by the proposal.
The addition of 12 months to the time available for performing
the necessary tasks by delaying the effective date of the
standards to the 1988 model year would provide adequate time
(30 months) for the performance of the tasks.
Two commenters provided statements to the effect that the
time necessary for implementation of the LOT NOx standards on
diesel vehicles was greater than that allowed by the proposal.
American Motors stated that it would have to add EGR and
particulate traps for simultaneous compliance with the NOx and
particulate standards of 1.2 g/mi and 0.26 g/mi respectively
for its LDDTiS and that the earliest possible date for
completion of this work was the 1988 model year. Nissan stated
-------
2-15
that the 1989 model year was the earliest possible date for
compliance with the NOx and particulate standard so as to allow
sufficient time for the devel-opment of new emission control
systems.
While both commenters integrated compliance with the
proposed 1.2 NOx standard for LDDTiS and compliance with the
1987 model year 0.26 particulate standard into their comments,
this integration is not'as important as could be inferred from
the comments since the requirements for the particulate
standard were promulgated in January 1984. Manufacturers have,
therefore, had ample time to plan and to initiate system
development and any tooling requirements associated with the
particulate standard. Timing requirements attributable to the
NOx standard are, therefore, the only requirements which need
to be addressed in analyzing these comments. Primary tasks
necessary for the application of EGR for the first tine (AMC
does not offer a diesel engine in their LDTs in California in
1985) or the application of a new EGR system design (Nissan)
'are: "1) overall system design which identifies exhaust
manifold changes to incorporate a source for the recirculated
gases and intake manifold changes to incorporate a point for
the introduction of the recirculated gases, 2) orders for
modified tooling for the manufacture of the redesigned
manifolds*, 3) build and test experimental systems at several
calibrations to establish one or more calibrations for use on
emission durability vehicles, 4) collect emission durability
data, 5) collect data from emission data vehicles, and 6) 'add
modified tooling to machine lines for the manufacturer of
redesigned intake and exhaust manifolds. Timing requirements
for tooling, to manufacturer EGR valves and plumbing, do not
enter into the overall timing considerations because these
parts can be expected to be supplied by existing facilities
(either vendor or r»3nufscturer owned). The timing requirement
estimates for the primary tasks in the critical timing path for
diesel LDTs are as follows:
Primary Task Time Required
Overall System Design - - 3-4 months
Develop Durability Vehicle Calibrations 5-7 months
Generate Emission Durability Data 11-12 months
Develop Final Calibrations 1-2 months
Run Data Vehicles 1 month
Complete Certification Process
With EPA and Add Modified Tooling . ' 1-2 months
TOTAL 22 - 28 months
Timing requirements for the procurement of a tooling
mpdification do not enter the critical path timing Line
since a complete new machine line can be delivered in 18
to 20 months.[2]
-------
2-16
Starting with a publication date of March 15, 1985 for the
final rule establishing the NOx standard, approximately 18-1/2
months would be available prior to the start of the 1S87 model
year in October of 1986. Since the time available for
performing the necessary tasks is less than the minimum
estimate for the requirements, it appears that there is merit
in the comments. The addition of 12 months to the time
available by delaying the effective date of the standards to
the 1988 model year would provide adequate time (30 months) for
the performance of the tasks.
4. Fuel Economy Effects of the Proposed Standards
For gasoline fueled LDTs, six commenters stated that the
proposed standards would result in a reduction in fuel
economy. These statements were made by American Motors,
Chrysler, Ford, General Motors, Nissan and Toyota. Only three
of the commenters (American Motors, Ford and General Motors)
however, provided numerical estimates of the effects of the
proposed standards on fuel economy. Ford indicated that a fuel
economy penalty of between 1 percent and 1.5 percent was
expected for heavy LDGTs at a NOx standard of 1.7 g/mi and
approximately a 2.5 percent penalty for heavy LDGTs at a 1.2
g/mi NOx standard. In addition, Ford stated that there would
be penalties in»the areas of driveability and performance.
GM stated that its LDGTs are exhibiting up to a 6 percent
fuel economy penalty when comparison is made between 1985 model
year Federal (2.3 g/mi NOx, 120,000 miles) and 1985 model year
California* LDGTs. GM continued its statement by saying that
the proposed NOx standard of 1.2 g/mi NOx with a useful life of
120,000 miles is more stringent than the California standard of
1.0 g/mi NOx with a useful life of 50,000 miles. In addition,
GM stated that while the application of three-way-catalyst
technology can be expected to improve fuel economy under a
constant NOx standard, it cannot be expected to either improve
fuel economy or to prevent a penalty under a more stringent
standard.
The comment by American Motors was similar to the GM
comment in that American Motors stated that its 1985 model year
1985 California NOx standards for LDTs up to 3999 . Ibs
equivalent inertia weight are 0.40 g/mi for 50,000 miles
with optional standards of 1.0 g/mi for 50,000 miles or
1.0 g/mi for 100,000, miles. For LDTs between 4000 and
5999 Ibs equivalent inertia weight the standard is 1.0
g/mi for 50,000 miles with an optional standard of 1.?
g/mi for 100,000 miles.
-------
2-17
California LDTs were exhibiting approximately 1 mpg* lower fuel
economy than its 1985 model year Federal' LDTs and that the 1.0
g/mi, 50,000 mile California standard was less stringent than
the proposed 1.2 g/mi, 120,000 mile Federal standard.
Since manufacturers had placed such emphasis in their
comments on the effects of the California NOx standards on fuel
economy, EPA assembled paired fuel economy data for 1985 model
year Federal and California specification LDTs from its
certification records. The criteria used in selecting paired
data were that the same engine, by manufacturer, had been
tested under the same dynamometer loading conditions in
vehicles equipped with the same transmission specification,
number of driven wheels (2 wheel drive and/or 4 wheel drive)
and N/V ratio. This information is shown in Table 2-7. The
information shown in Table 2-7 is subdivided into three groups,
those gasoline LDTs employing the same technology for
compliance with the Federal and California standards, those
using -different technologies and diesel .LDTs. Because
California NOx emission standards include several options and
the specific option used was not always clearly defined in the
records, exact distinction between the vehicle emission levels
and the useful life requirement of the standard was not
achievable. Distinction was possible, however, between
vehicles certified to the 0.40 g/mi, the 1.0 g/mi and the 1.5
g/mi standards and is shown on Table 2-8.
EPA's overall assessment of the California versus Federal
comparisons is that they are of limited use in making precise
conclusions about the effects to be expected from the new LOT
standards. This is first of all due to the fact that the
California standards are more stringent than the Federal
standards, 'resulting in somewhat lower emission levels than,
will the Federal standards. Under this situation, there will
be a somewhat greater impact on fuel economy associated with
the California levels. In addition, consideration must be
given to the fact that the Federal standards will not apply
until the 1988 model year. This means that time would be
available for improvements aimed at overcoming any fuel economy
penalties which might currently exist. Lastly, it must be born
in mind that California standards apply to only a small
fraction of any manufacturers' total LOT sales. Therefore, the
manufacturers can be expected to adopt the lowest initial cost
approach to compliance with a relatively small concern over
fuel economy effects. This will not be the case in the longer
term, when the entire LOT fleet is -affected.
One mpg corresponds to between a 4 percent arid a 6 percent
fuel economy reduction for American Motors when based on
comparisons to the highest "highway" fuel economy estimate
or to the lowest "city" fuel economy estimate.
-------
2-18
Table 2-8
L985 Model "tear Light-Duty Truck Fuel Economy
Federal vs. California Paired Data*
Fuel Economy
Technology
Mfr.
Gasoline, Same
RAC
RAC
RAC
Chrysler***
Ford
Ford
Ford
Ford
QVJ *** - -
04
Nissan
Mitsubishi
Toyota ***
Toyota ***
VW
Engine**
Technology
2.5LUO)
4.2L(7)
5.9L(1)
5.2LC2)
2.3L(4)
2.8L(7)
4.9L(6)
5.0L(2)
2.5L(19)
4.3L(2)
2.4L(3)
2.6L(13)
2.0U4)
2.4U6)
1.9L(2)
Fed.
3CL
3CL
3WY
3CL-KJX
3CL
3CL40X
3CL-K3X
3CL-KK
3CL
3CL-+OX
3CL
3CL
3CL
3CL
3CL
Calif.
3CL
3CL
3WY
3CL-KK
3CL
3CL-K3X
3CL-K3X
3CL-K3X
3CL -
3CL-K3X
3CL
3CL
3CL
3CL
3CL
Combined MPG %
Fed.
23.4
20.6
13.8
14.8
29.8
22.6
17.2
17.1
27,0
22.6
23.3
23.2
27.0
25.4
20.1
Calif.
22.5
20.6
13.8
14.8
29.4
21.4
16.7
16.4
27.0
22.0
23.1
22.4
27.0
25.4
18.9
Change
-3.8
0
0
0
-1.3
-5.3
-2.9
-4.1
0
-2.6
-0.1
-3.4
0
0
-6.0
NOx Emissions
Fed.
1.04
1.30
1.41
0.74
0.49
1.51
1.26
0.83
0.30
0.53
1.46
1.27
0.11
0.14
0.92
Calif.
0.55
0.70
0.55
0.74
0.26
0.66
0.68
0.54
- 0.30
0.54
0.52
0.64
0.11
0.14
0.39
Gasoline, Different Technology
RAC
Chrysler
04 - ...:
G4
04 ,. - " '
Isuzu
Toyota
Fuji
Diesel
GM
2.8L(l)
5.2L(1)
2.8L(7)
4.3L(2)
5.7L(1) '
1.9L(7)
2.4L(4)
1.8L(6)
6.2L(7)
OX
OX
OX
OX
OX
OX
OX
OX
BGR
3CL
3CMCX
3CL .
3CL
3CL
3CL
3CL
3CL-KDX
Elec EGR
20.5
15.5
22.8
20.1
13.9
27.5
28.8
27.3
23.5
22.1
15.2
24.7
19.5
14.2
26.8
27.2
28.6
-
23.3
7.8
-1.9
8.3
-3.0
2.2
-2.6
-5.6
4.8
-0.1
1.80
1.81
1.44
1.57
1.80
1.77
1.62
1.78
1.59
0.76
0.85
'0.65
0.47
0.80
""0.41
0.40
0.20
1.28
Calif.
Std.
1.0
1.0
1.0
1.0
0.4
.0
.0
**
***
1.0
0.4
1.0
1.0
1.0
0'.4
0.4
0.4
1.0
1.0
1.0
1.0
1.0
1.0
0.4
0.4
1.5
Data pairing requirements: equal engine displacement, transmission, N/V and inertia
weight.
Number in ( ) following engine size identifies the number of engine pairs used in
calculating the mean fuel economy values shown.
No changes occur because this is a 50 state vehicle. Use of the same engine for Federal
and California versions implies minimal fuel economy impact associated with the required
low NOx level.
-------
2-19
Regardless of the above caveats, instructive conclusions
can be drawn from the v Ca lif o"r_nia data. For those vehicles
already having three-way technology on the Federal versions,
there is no clear pattern of fuel economy change between the
Federal and California versions. The data show that in some
cases there are marked reductions in NOx emissions with little
or no fuel economy impact. On the other hand, some vehicles
exhibit significant penalties. However, the higher penalties
are generally associated with vehicles having California NOx
levels well below those needed for compliance with the 1.2/1.7
Federal standard. In any event, it has already been noted that
vehicles already equipped with three-way technology are largely
already in compliance with the 1.2/1.7 standards. Therefore,
no changes will be required of these vehicles and no fuel
economy impact will occur.
For those systems configured with oxidation catalysts on
the Federal version, the data of Table 2-8 confirms EPA's
-analysis -from the Draft RIA. All cases switching from
oxidation catalysts to three-way catalysts, except for some
certifying to unnecessarily low NOx levels, show a significant
gain in fuel economy.
Overall, EPA draws the .following conclusions regarding
fuel economy effects on gasoline-fueled LDTs of the new LDT
standards. First, for those vehicles already employing
three-way technology, compliance or near-compliance is already
widespread. Therefore, no fuel economy impact will result from
the new standards. Second, for those vehicles switching from
oxidation catalysts .to three-way systems, a significant
improvement in fuel economy should result from the new
'technology at the NOx levels associated with the Federal
standard. In total, there will probably be some small fuel
economy gain associated with the new standards. Since the
amount cannot be precisely quantified at this time, no specific
benefit will be included in the economic analyses of the new
standards.
In the case of LDDTs, GM is the only commenter to comment
on the fuel economy effects of the proposed standards. The
comment provided by GM was directed to their 6.2 liter engine
and indicated that GM expected a fuel economy penalty as a
result of the 'proposed 1.7 g/mi standard which would be greater
than the 5 percent which they are experiencing under the 1985
model year California standard.
Inspection of the NOx emission levels for the California
6.2 liter engine (Table 2-8) shows that the engine is certified
to a 1.5 g/mi standard in conjunction with the particulate
standard of 0.4 g/mi. At these California standard levels, the
change, in fuel economy relative to the Federal standard of 2.3
g/mi, NOx and 0.60 g/mi particulate is 0.1 percent, i.e. there
-------
2-20
is essentially no difference between the fuel economy values
developed for the 6.2 liter GM engine under 1985 model year
Federal and California standards. Since the proposed Federal
standard applicable to this engine (1.7 g/mi) is not
numerically as stringent as the California standard, EPA sees
no basis for the comment provided by GM. The conclusion which
can be drawn from the information shown in Table 2-8 is that
there should be no fuel economy effect on the 6.2 liter GM as a
result of the proposed standard.
5. Other Comments
Other comments pertaining to the proposed LOT "NOx
standards were provided in the areas of the factors to be used
in distinguishing between LDTts and LDT2s, the
comparability between the 1.2 g/mi proposed standard and the
1.0 g/mi LDV standard and the proposed high altitude standards
for NOx, idle CO and particulate.
Since these comments are fully addressed in the preamble
to the final rule, they are not analyzed here. EPA agrees with
the need to correct the discriminator between LDTts and
LDT2s. However, none of the comments in the other areas
substantiate a need for changes. Interested readers 'are
referred to the preamble for further details on EPA's response
to these comments.
C. Conclusions
As a result of the proceeding analyses of the comments
-provided in response to the NPRM, it is EPA's conclusion that
the proposed NOx standards of 1.2 g/mi for LOT t s and 1.7 g/mi
for LDT2s are technologically feasible for the respective
groups of LDTs. The technologies expected to be used in
complying with these emission standard levels will center on
the use of three-way catalyst technology with closed-loop fuel
control in the case of gasoline-fueled LDTs and on the use of
EGR in the case of diesel-fueled LDTs.
Analysis of the comments has led EPA to conclude that the
time required (leadtime) for implementation of the necessary
technologies is greater than that which would be available with
an implementation date of the 1987 model year. A one-year
delay in the implementation date of the standards to the .1988
model year would, however, provide sufficient leadtime.
Analysis of the comments provided on the fuel economy
effects of the proposed standards has lead EPA to conclude that
on average, those LDGTs which are already equipped with
three-way catalyst technology will experience little or no
reduction in fuel economy while those LDGTs which are converted
-------
2-21
from oxidation to three-way catalyst technology are expected to
experience increases in fuel economy. In the case of LDDTs,
the expectation is that there will be no measurable change in
fuel economy resulting from the NOx standards. For the total
fleet- of LDTs, the effect of the proposed standards on fuel
economy is expected to be near zero but with the potential for
some improvement resulting from those LDGTs which adopt
three-way catalyst technology.
III. Heavy Duty Gasoline Engines (HDGEs)
A. Synopsis of NPRM Analysis
The NPRM analysis[1] examined the feasibility of the
proposed 1987 6.0 g/BHP-hr and the 1990 4.0 g/BHP-hr NOx
emissions standards for HDGEs. The analyses for each standard
began with the identification of the appropriate low-mileage
target values. Current HDGE emission levels were then
.discussed as part of the analysis of the 1987 standards, as
well as the effects of leadtime constraints and available
emission control technologies. The analysis for the 1990
standard considered the likelihood of new and more refined
emission control technologies. A summary of the NPRM analysis
follows.
IT. ( "
1. 6.0 g/BHP-hr NOx Standard
The factors considered in estimating the low-mileage
emission target were the additive deterioration factor and the
production variability factor. The additive deterioration
factor (DF) was developed from 1983 model year HDGE
certification data and was found to be zero. The NPRM's
production variability factor of 1.2 was the mean of two
estimates previously provided by Ford and GM in response to an
earlier rulemaking. These two factors were employed in the
following equation to develop the low-mileage emission targec
of 5.0 g/BHP-hr.
Emission Standard - Deterioration Factor
Low Mileage"Target «
Production Variability
The second step in the analysis was the identification of
the reductions in emissions required to meet the target level.
This was accomplished by a comparison of the low-mileage target
and the most up-to-date information on- low-mileage emission
levels actually being achieved. The most up-to-date data
available were those developed from prototype 1985 model year
HDGEs. As a result of the comparison it was found that four of
eleven engine families could presently (1985) comply with the
low-mileage target. The average reduction necessary for
compliance with the standard by the remaining seven families
-------
2-22
was 15 percent. The greatest reduction necessary for
compliance by an engine family was 34 percent and the lowest
reduction was 7 percent.
In considering leadtime, the analysis noted that some
development testing had already been performed and further
development testing would be initiated during the course of the
rulemaking. Still, somewhat less than the equivalent of two
years leadtime was determined to be available for NOx control
development, thus, precluding the availability of major engine
or hardware changes for production.
The final step in the analysis was the identification- of
technologies which could provide the reductions in NOx
emissions necessary for compliance. Three potential
technologies were identified: ignition timing retard, fuel
enrichment of the air-fuel charge delivered to the engine and
EGR. Ignition timing retard as the sole method of achieving
compliance was judged to be unacceptable since it would result
in a relatively large fuel economy penalty. Fuel enrichment
was also judged to be undesirable since it would negatively
impact compliance with both the HC and the CO standards as well
as causing a reduction in fuel economy. Increased EGR,
possibly coupled with a small amount of timing retard, -was"
judged to be the approach which would most probably be employed
by manufacturers since the necessary reduction in NOx could be
achieved with an insignificant effect on fuel economy.
The analysis concluded that, based on information then
available, and considering the relatively modest reductions
necessary for only a fraction of the fleet and based on the
availability of well understood NOx control technologies for
gasoline-fueled engines, a 1987 NOx standard of 6.0 g/BHP-hr
was feasible for HDGEs.
2. 4.0 g/BHP-hr NOx Standard
The low-mileage emission target for the 4.0 g/BHP-hr
standard was developed using the same procedure as that used
for the 6.0 g/BHP-hr standard. The same values for the
deterioration factor and the production variability factor were
also used because compliance with the 4.0 g/BHP-hr standard was
expected to be achievable without the use of reduction catalyst
technology. The low-mileage emission target developed by this
procedure was 3.3 g/BHP-hr.
The reductions from 1985 model year prototype levels
necessary for compliance with the 4.0 g/BHP-hr standard were
estimated once the low-mileage target level had been
identified. The .average reduction necessary for compliance was
found to be 39 percent with the greatest reduction being 57
percent and the least reduction being 3 percent.
-------
2-23
At the level of emission control required for compliance
with a 4.0 g/BHP-hr, it was concluded that emission control
technologies beyond those required for compliance with the 6.0
g/BHP-hr standard {i.e., standard EGR) could be required to
avoid significant performance and fuel economy penalties. The
technologies identified as being the most probable for use were
increased EGR rates with improved controls and "fast-burn"
combustion chamber design, coupled with probable use of
electronic control to optimize fuel metering and ignition
timing.
- With respect to leadtime, the adoption and demonstration
of these control technologies were considered, at that time to
be feasible for 1990, based on the fact that prototype engines
were already approaching the design target and considerable
experience was directly transferable from work in light-duty
vehicle and light-duty truck NOx control.
B. Summary and Analysis of Comments ' " -----
The Agency received comments on its NPRM analysis from the
three manufacturers of heavy-duty gasoline engines: Chrysler,
Ford, and General Motors. Their comments on the 6.0 g/BHP-hr
standard are examined first followed by an analysis of those
pertaining to the 4.0 g/BHP-hr standard.
As will be seen in the next section on HDDEs, the 6.0
g/BHP-hr NOx standard will not be feasible for HDDEs until
1988. Thus, this implementation date will be assumed here, as
well. Also, the 4.0 g/BHP-hr standard was found not to be
feasible for HDDEs by 1990. However, a 5.0 g/BHP-hr standard
appears to be feasible for 1991. Thus, this will be the
second-stage NOx standard considered here for HDGEs.
1. 6.0 g/BHP-hr NOx Standard
None of the manufacturers disagreed with the low-mileage
target level of 5.0 g/BHP-hr, nor with the low-mileage
prototype data presented in the NPRM analysis. With respect to
the availability of control technology and leadtime, two of the
three manufacturers were generally in agreement with the
conclusions reached in the NPRM analysis. In their submittals,
both Ford and Chrysler stated that they could meet the proposed
6.0 g/BHP-hr standard; Ford in 1987 and Chrysler in 1988. GM
stated that this standard should be feasible for its HDG
vehicles above 14,000 Ibs. GVW, "but would result in a 1.5
percent fuel economy penalty; GM added that for its engines
used in 8,500-14,000 Ib. GVW vehicles, it did not believe that
the proposed standard was feasible in conjunction with the 1987
model year 1.1/14.4 g/BHP-hr HC/CO standards. As an
alternative to the 6.0 g/BHP-hr level, GM recommended a HDE NOx
-------
2-24
standard of 8.0 g/BHP-hr. However, this was due to GM's
continued belief that catalyst*technology is still not feasible
for these engines, and not on an inability to meet the NOx
standard, per se. Thus, given the fact that the initial NOx
standard is being delayed to 1988 for HDDEs, all three
manufacturers essentially agree that the 6.0 g/BHP-hr standard
is feasible for HDGEs.
With respect to the technology needed to comply with this
standard, both Ford and GM disagreed with the NPRM's 'assessment
that increased EGR, possibly coupled with a small amount of
timing retard, was sufficient and the most likely approach to
be employed. According to Ford, more than just increased "EGR
and ignition timing retard are required in order to comply with
the regulations while maintaining the fuel economy, performance
and driveability of Ford's heavy-duty vehicles. In its
confidential comments, Ford listed the control techniques it is
planning to incorporate in order to meet a 6.0 g/BHP-hr
standard.
GM also criticized the Agency's assessment of EGR as a
control technique because of the fuel economy penalty resulting
from increased EGR. However, unlike Ford, GM did not believe
that alternative techniques were available for its HDGEs.' GM
supplied data taken on a 1985 350-4 V8 prototype engine that
showed a 1.5 percent fuel penalty resulting from an increase in
the EGR in order to comply with the proposed standard. Also,
both recalibration of the air-fuel ratio and retarded ignition
timing were found to be unacceptable by GM for the same basic
reasons as identified in the NPRM. Chrysler did not comment on
the technology needed for its engines to comply with a 6.0
g/BHP-hr standard.
Neither Ford nor GM presented sufficient justification for
their projections of technology requirements to allow them to
be objectively critiqued here. However, an analysis of 1985
Federal HDGE certification data confirms the conclusion of the
NPRM that EGR is basically capable of providing the degree of
control necessary to meet the 6.0 g/BHP-hr standard (see Table
2-9). Two engines, a 7.5L Ford and a 7.4L GM, are already
being certified at NOx levels of 4.2 and 4.5 g/BHP-hr,
respectively. The only significant difference between these
engines and those at higher NOx levels appears to be increased
EGR and recalibrated engine parameters (i.e., timing, secondary
air rates, etc.). Thus, more significant changes should not be
required for most HDGEs. As roughly one-third of all 1985
prototype HDGEs were able to comply with a 6.0 g/BHP-hr NOx
standard and another one-third were within 25 percent of the
standard, these engines should require no more than increased
EGR rates plus recaLibration. However, as described in the
Draft RIA, the NOx levels of some of the engines were well
-------
2-25
•teble 2-9
1985 HDGE Federal Certification Results (g/BHP-hr)
Manufacturer Displacement Emission Control NOx (DF) HC (DF) CO (DF)
Ford 4.9 EGR-Air 8.49(.0l) 1.82(0.0) 15.65(.00)
14.93(.00)
30.65(.ll)
31.62(.07)
14.01(.03)
14.45(.00)
23.4K.OO)
25.77(.00)
29.08(5.03)
25.32(5.03)
27.48(5.03)
Chrysler 5.9L BGR-Air 7.7K.07) .65(.00) 18.73(.00)
Displacement Emission Control
4.9
5.8
7.5
- 4.8
5.7
.
6.0
7.0
7.4
5.9L
E3R-Air
B3R-Air
EGR-Air
B3R-Air
B3R-Air
Air
B3R-Air
EGR-Air
EGR-Air
BGR-Air
EGR-Air
BGR-Air
NOx (DF)
8.49(.0l)
6.96(.00)
8.24(.00)
6.66(.00)
4.2K.OO)
7.03(.00)
8.33(.05)
5.82(.05)
7.62(.00)
7.72(.00)
4.5K.OO)
7.7K.07)
HC (DF)
1.82(0.0)
1.66(.0l)
1.78(.00)
.96(.00)
.41(.00)
.97(.00)
1.47 (.02)
1.3H.02)
1.29(.25)
1.23(.25)
.68{.25)
.65(.00)
-------
2-26
above the design target of 5.0 g/BHP-hr (i.e., more than 25
percent). Compliance by these "engines, which represent roughly
half of those not already in compliance with the 6.0 g/BHP-hr
standard, may require more significant modification to avoid
impacting either HC/CO emissions or fuel economy. These
modifications were among those identified in the NPRM analysis
for the 4.0 g/BHP-hr standard and include modifications to the
combustion chamber, the intake manifold, the secondary air
system, and the camshaft. As discussed in Chapter 3, these
changes may require some retooling, but, given their nature and
the comments of Ford and Chrysler, leadtime should not be
affected.
The certification levels shown in Table 2-9 are generally
higher than those of the prototype engines described in the
NPRM. This does not necessarily imply that the levels of the
prototype engines were not in the end achievable. The current
NOx standard puts little, if any, real pressure on HDGE NOx
emissions, so the higher certification results probably
involved recalibration to higher NOx levels. Thus, this does
not negate the potential to achieve the lower NOx levels with
the two sets of engine modifications described above.
Also to be noted from Table 2-9 is the positive
relationship between HC and NOx emissions (i.e., HC emissions
decrease as NOx emissions decrease). This is not to say that
EGR decreases HC emissions, but that other engine parameters,
such as the secondary air injection rate, can be adjusted to
eliminate any adverse effect of EGR on HC emissions. This
positive relationship is present even at the two lowest NOx
levels of 4.2 and 4.5 g/BHP-hr.
With respect to fuel economy, GM argued for a 1.5 percent
penalty, while the other two manufacturers did not comment on
the NPRM's projection of no penalty. GM based its judgment on
testing of a single engine with varying EGR rate. It was not
clear from the information presented if BSFC was optimized at
each EGR rate, or if EGR was simply increased. No actual data
nor engine calibrations were presented. Thus, the GM
projection cannot be evaluated against the other three
projections of no penalty. Thus, the conclusion of the NPRM
will be carried forward here, that of no fuel penalty.
In summary, essentially all three manufacturers of HDGEs
are in agreement with the Agency's conclusion that a 6.0
g/BHP-hr standard is feasible for 1988 model year HDGEs. This
standard is obtainable for HDGEs within the available leadti.T.e
constraints, and should result in no undue fuel economy,
performance, or driveability penalties.
-------
2-27
2. 5.0 g/BHP-hr NOx Standard
In comments on the proposed 1990 model year 4.0 g/BHP-hr
NOx standard, the manufacturers uniformly termed this standard
infeasible. Chrysler did not believe that the technology which
will be available by the 1990 model year will be capable of
achieving the 4.0 g/BHP-hr standard. Thus, Chrysler felt that
the Agency did not realistically assess the prospects that the
necessary control technology could be produced in time to
assure compliance.
«
GM reported that its effort to reduce NOx emissions from
HDGEs used in trucks above 14,000 Ibs. GVW from current levels
to the level required to comply with the proposed standard, HC
.emissions were doubled and fuel consumption was increased by
about 6 percent. Thus, GM believed that the 4.0 g/BHP-hr NOx
standard was not feasible because it would prevent compliance
with the 1.9 g/BHP-hr non-catalyst HC standard for 1987 model
year heavy HDGEs; also, the fuel assumption penalty was
unacceptable.
Ford contended that EPA erred in its technological
feasibility assessment of the control methods required to meet
the standard. Ford was convinced that in order to reduce NOx
emissions to the 4.0 g/BHP-hr level, a three-way-catalyst was
required. According to Ford, a three-way catalyst is not
capable of operating under the high-temperature conditions
encountered by Class IIB, III, or VI heavy-duty trucks.
Therefore, it determined that the 4.0 g/BHP-hr standard was not
feasible. Ford also questioned EPA's analysis of fast-burn
technology "as a control method; Ford believed that the burn
rates of the fast-burn cylinder heads described in the NPRM
analysis will not be significantly different than a
conventional head at tne high speed and load conditions cf the
heavy-duty transient test cycle, thus making no allowance for
further EGR optimization.
Since the 4.0 g/BHP-hr standard is no longer being
considered for HDGEs the above comments pertaining to the 4.0
g/BHP-hr standard must be analyzed with respect to a 5.0
g/BHP-hr level. However, little detailed technical analysis
was provided by the commenters to contribute to a detailed
assessment of either a 4.0 or 5.0 g/BHP-hr standard. Thus, the
analysis here will rely on the analysis performed for the NPRM
and 1985 certification data. Further, an adoption of a 5.0
g/BHP-hr NOx standard should. mitigate many of the
manufacturers' concerns.
The NPRM analysis stated that the low-mileage target for a
5.0 g/BHP-hr NOx standard would be 4.2 g/BHP-hr. Based on
1985-1987 prototype data, that analysis also showed two engines
-------
2-28
already to be below this level and the remainder requiring an
average 30 percent reduction in NOx emissions. Available 1985
Federal certification data (Table 2-9) basically confirm this.
One engine is at the 4.2 g/BHP-hr target, while another is just
slightly above this at 4.5 g/BHP-hr; these levels are being
achieved essentially with EGR and minor engine recalibration.
The remaining 1985 engines require somewhat more than a 30
percent reduction on average. However, this is not significant
since the current 10.6 g/BHP-hr NOx standard puts no pressure
on NOx emissions, and, therefore, there was no guarantee that
the low NOx levels achieved by prototype engines would appear
in certification. Given the fact that two engines in
production already essentially meet the low-mileage target "and
a third protoype engine also met this level over a year ago, it
is difficult to argue that this level will not be feasible six
years hence. This is especially true given the general
homogeneity of HDGE technology, which stands in stark contrast
to the heterogeneous HDDE technology. The technologies
discussed in the NPRM are applicable to any HDGE. Thus, the
5.0 g/BHP-hr NOx standard should be feasible for HDGEs.
This standard will require control technology similar to
that required for the 6.0 g/BHP-hr standard (i.e., combustion-
chamber modifications, improvements to the intake manifold,'the
secondary air system and the camshaft). However, because a
greater level of NOx reduction is required to reach 5.0
g/BHP-hr, a larger percentage of the fleet will require these
hardware modifications in addition to increased EGR rates and
recalibrations; burn rate improvements, as described in the
NPRM analysis, may also be required as a control technology.
Since roughly 15 percent of the current HDGEs of Table 2-9
essentially comply with the 5.0 g/BHP-hr standard without these
hardware modifications and assuming NOx averaging to be
available, it is estimated that roughly one-third of the
remainder will be able to do so as well. Therefore, of the
approximately 85 percent of the fleet requiring any additional
control, about two-thirds will require the hardware changes
described above, in addition to increased EGR and engine
recalibration.
Although the 5.0 g/BHP-hr should be feasible via
engine-related changes as detailed above, this does not rule
out the possibility that manufacturers will decided to apply
three-way catalyst technology to meet the standard. Class IIB
and III HDGVs will be equipped with oxidation catalysts in 1987
to comply with the HC/CO emission standards and their LDGT
counterparts will likely be equipped with closed-loop,
three-way catalyst technology. Thus, the step to three-way
catalyst may be considered by some manufacturers. However,
such a change -is not likely, since manufacturers have
repeatedly emphasized to the Agency their position that
-------
2-29
significant questions of feasibility exist for three-way
catalysts in the heavy-duty environment. It was for reasons
such as these, and their associated cost impacts, that EPA
chose not to propose a three-way catalyst based standard for
HDGEs 'in the proposal.
If done, application of three-way systems would involve
increased initial vehicle cost. However, fuel economy and
performance should improve beyond current levels, as indicated
in Section II above for LDGTs. Otherwise, no substantial
adverse impact on fuel economy, performance or driveability is
expected, due to the substantial leadtime involved and the
hardware modifications available. "Thus, a manufacturer would
only be expected to apply three-way catalysts if it resulted in
a net cost-benefit improvement with respect to its profits and
consumer satisfaction.
C. Conclusions
The following conclusions result from the preceeding
analysis of the comments provided on the technological
feasibility of the proposed standards and from the draft
regulatory analysis performed in support of this rulemaking.
A NOx standard of 6.0 g/BHP-hr should be feasible for 1988
'model year heavy-duty gasoline,^ engines. Roughly one-third of
all HDGEs are already in compliance with this standard without
any hardware modifications from their higher NOx counterparts.
One-half of the remainder will require only increased EGR rates
,and engine recalibration to comply. The other half will
require hardware modifications in addition to increased EGR and
recalibration. Complying with a 6.0 g/BHP-hr standard should
not result in undue fuel economy, performance, or driveability
penalties for HDGEs.
A NOx standard of 5.0 g/BHP-hr should be feasible for 1991
model year heavy-duty gasoline engines. Roughly 15 percent of
current HDGEs are already in compliance with this standard
"without any hardware modifications from their higher NOx
counterparts. Assuming that NOx averaging will be available,
roughly two-thirds of the remainder will require only increased
EGR rates and engine recalibration to comply. The other
one-third will require minor hardware modifications in addition
to increased EGR and recalibration. This increased application
of control technology should avoid any measurable performance
or fuel economy penalties at the 5.0 g/BHP-hr standard level.
-------
2-30
IV. Heavy-Duty Diesel Engines (HDDEs)
In developing the proposed emission standards for HDDEs,
the NPRM analysis[l] treated the process in two distinct
stages. In the first stage, the focus was on the
identification of achievable emission levels for NOx and
particulate emissions in the near-term (1987). In the second
stage, the focus was on levels achievable in the mid-term
(1990). The second stage of the development process included
the evaluation of feasible engine-out NOx and particulate
emission levels as well as the feasibility of trap technology.
The identification of feasible engine-out NOx and particulate
levels are clearly related; consequently, they are discussed
together and the analysis of trap feasibility and associated
particulate standard levels is treated separately. Thus, the
near-term NOx and particulate standards are examined first,
followed by the mid-term NOx and non-trap particulate standards
and then the trap-based particulate standards.
A. Synopsis of NPRM Analysis
1. Near-Term NOx and Particulate Standards
The NPRM draft regulatory analysis[l] of the technological
feasibility of the proposed 1987 NOx and .particulate standards,
6.0 g/BHP-hr NOx and 0.60 g/BHP-hr particulate, consisted of
five steps and is summarized as follows. The first step was
the identification of NOx and particulate emission levels from
current engines. These data were broken down by HDDE subclass
(light (LHDDE), medium (MHDDE), and heavy (HHDDE)), because of
the technological differences in engine designs between these
subclasses. NOx emission levels were obtained from both
Federal and California certification data. However, as
particulate emissions are not currently regulated, these data
had to be gathered from a variety of sources.
The second step in the analysis was the determination of
the low mileage emission targets and the amount of emission
reduction necessary for compliance with the proposed
standards. The identification of the target level was
performed according to the same basic methodology described
above for LDTs and HDGEs. With respect to the amount of
emission reduction required, HDDEs were divided into two
groups: indirect injection (IDI) and direct injection (DI)
engines. In the case of IDI engines, (engines manufactured by
GM -and IH), it was concluded that available transient test data
on the GM engine showed that it could already comply with the
proposed standards. Steady state data on the IH engine
strongly suggested that it also could comply. As for the DI
engines, which constitute the majority of the HDDE families,
all exhibited higher NOx and particulate levels than was the
case for the IDI engines. Substantial differences between
-------
2-31
various DI engine configurations were identified (naturally
aspirated, turbocharged or turbbcharged wifeh aftercooling).
T,he third step in the HDDE analysis on engine-out
emissions was the identification of the technologies which
could provide the necessary emission reduction necessary for
compliance with the proposed standards. The analysis for
LHDDEs (roughly equivalent to IDI engines) was fairly
straightforward, since the available data indicated that these
engines were already at or below the 1987 standards. The MHDDE
and HHDDE (roughly equivalent to DI engines) analysis was more
complex and began with an estimation of short-term BSFC
improvements, since reductions in fuel burned translate
.directly to reductions in NOx and particulate emissions. The
analysis then moved on to an assessment of the effectiveness of
various techniques to directly control NOx and particulate
emissions, including injection timing retard and aftercooling.
.On the ..basis of the wide variation in engine design
configurations present and the disparity in emissions, "it
appeared that each manufacturer, for each of its engines, could
adopt multiple emission control strategies for compliance with
emission standards.
The fourth step was an assessment of the effect of the
proposed 1987 standards on HC emissions and fuel economy. In
estimating the fuel economy effects of the proposed standards,
EPA considered estimates provided by manufacturers as well as
estimates developed by the National Academy of Sciences (NAS)
'(page 2-76 in Reference 1). The resulting estimated effect on
fuel economy was for up to a two percent reduction initially
diminishing to zero by the third year of the proposed standards
-(6.0/0.60) .
Leadtime was the final step. Since the emission control
strategies expected to be used in complying with the proposed
standards involved recalibrations of injection timing,
modification of aftercooling and/or the addition of
aftercooling on some engines, the leadtime required for the
implementation of the proposed standards was considered to be
within the time proposed for implementation.
2. Mid-Term NOx and Non-Trap Particulate Standards
In the NPRM analysis, the assessment of the feasibility of
the 1990 NOx and non-trap particulate standards was performed
in three steps and is reviewed as "follows. The initial step of
the 1987 analysis, the determination^ of current emission
levels, did not have to be repeated, since it could be assumed
that all engines would be at the design targets necessary to
meet the 1987 standard. Thus, the first step developed the
target- levels for the proposed 4.0/0.40 standards; the sane
methodology as had been employed for the 6.0/0.60 proposed
-------
2-32
standards was used. The target levels developed were 3.2-3.6
g/BHP-hr NOx and 0.30-0.33 g/BH'P-hr particulate.
Next the analysis assessed the effectiveness of various
control techniques. Technologies projected to be required for
compliance with 4.0/0.40 engine-out emissions standards were
broader than those anticipated for the 6.0/0.60 proposed
standards and included the following; additional injection
timing retard, advanced aftercooling designs, improved
combustion chambers, high pressure fuel injection, exhaust gas
recirculation, electronic controls, in-cylinder heat retention
and fuel modification. In light of the wide variation which
exists between specific diesel engines, it was anticipated that
manufacturers would, on an engine specific basis, select the
combinations of these technologies most appropriate for each
engine.
Finally in the third step, the effect of the proposed
standards on fuel economy was examined. In the short term,
i.e. immediately following the effective date of the proposed
standards, the projected effects of the 4.0/0.40 proposed
standards on fuel economy was for a 1-2 percent penalty which
should be eliminated with time.
3. Trap-Based Particulate Standards
In the NPRM analysis[1] of the feasibility of particulate
trap-oxidizers for heavy-duty diesel use, EPA determined that
traps would be feasible for 1990 model year HDDEs. Due to the
limited amount of available HO trap development data, the
analysis first examined light-duty trap status and then
considered the degree of additional development effort required
by the heavy-duty industry. As a result of this and the
ongoing research and development data, EPA concluded that
trap-oxidizers would be available to permit compliance with a
HDDE trap-based particulate standard; the standard level was
also calculated in this analysis. The following will synopsize
the four steps of the NPRM analysis: LD trap status; LD/HD
differences; HDD trap status; and emission levels.
Based on past EPA analyses and a contracted study [3,4,5],
the Agency concluded that light-duty trap technology was at a
very advanced stage of development and light-duty trap
oxidizers would be technically feasible no later than the 1987
model year. The findings that traps were feasible for 1987
model year LDVs was also based on Daimler-Benz's plans to
certify a trap-equipped vehicle to meet California's 1985
emission standards. Although1 there were still unresolved
problems associated with some trap systems (e.g., introduction
of a fuel additive to the fuel to induce regeneration, the
development of a fully automated positive regeneration system,
and the occurrence of increased sulfate emissions from
-------
2-33
catalyzed traps), EPA believed that the other manufacturers
were not far behind Daimler-Benz's trap development and
compliance was possible for 1987 LDV use.
The second step in the analysis examined the applicability
of light-duty trap technology to heavy-duty engines, concluding
that there was nothing preventing the adaptation of light-duty
technology, with additional development, to heavy-duty usage.
The further advanced light-duty trap technology formed the
basis for the development of similar technology for the
heavy-duty diesel engine industry. However, conditions
- specific to the HDDE environment were identified which must be
considered in the design of heavy-duty trap, oxidizers. The
major light-duty/heavy-duty differences included: engine size
and load factor; operating conditions and temperatures; the
useful life of the engine; and ash accumulation. Although
considerable development effort was found to be required of the
heavy-duty industry in the adaptation of light-duty trap
technology for heavy-duty use, EPA did not consider the
problems to be without engineering solutions.
The analysis continued with a survey of the ongoing
heavy-duty trap research and development. The Agency found a
definite lack of data from -the HDDE industry, noting the
difference between LD trap progress, where the LD industry has
had to work towards a trap-based standard, and HD trap
progress, where the HD industry has not had that incentive.
• The limited development work was primarily focused on trap
regeneration and its control. (Trap type, for the most part a
direct derivative of light-duty design, was not considered a
major obstacle, although some design effort in this area
remained). Regeneration methods being evaluated included, but
-were not limited tc, burners, fuel additives and catalyzed
traps. Development of an automatic regeneration control system
appeared to be the next major step. EPA realized that traps
were not at the time a viable particulate control, but the
Agency firmly believed that, with industry's vigorous pursuit
of a trap-oxidizer system, traps would be achievable for HDDEs
" by 1990.
The final step identified a feasible standard for
trap-equipped heavy-duty diesel engines. The proposed trap
standard was dependent on the following factors: the
engine-out design target level, the deterioration factor (DF),
the SEA adjustment factor and the trap efficiency. The target
level, SEA adjustment factor and DF for the engine-out emission
level of 0.60 g/BHP-hr were determined in the non-trap standard
section of the analysis. The 1.0 DF for traps was based on 1DD
particulate emission tests of over 50,000 miles that resulted
in no significant deterioration. The final and most variaoie
factor-, the trap efficiency, ranged from 50 to greater than 90
percent, dependent on trap type. The Agency determined that
-------
2-34
with 80 percent efficient traps, HDDEs could comply with a 0.25
g/BHP-hr standard; with averaging, approximately 70 percent of
the HDDEs would require traps. If essentially all vehicles are
equipped with 90 percent efficient traps then a 0.10 g/BHP-hr
standard was determined to be feasible.
The technology feasibility analysis concluded that there
appeared to be sufficient time for the manufacturers to design,
develop, and prepare trap-oxidizers for 1990 model year HDDEs.
This followed from the fact that traps will be in production on
many light-duty diesels no later than 1987. The five years of
leadtime between mid-1984 and late 1989 were found to allow
adequate time for the additional design effort required -for
HDDE modifications.
B. Summary and Analysis of Comments
1. Near-Term NOx and Particulate Standards
The proposed 6.0/0.60 standards for 1987 represented an
attempt by EPA to obtain meaningful, yet balanced, reductions
in both NOx and particulate in the near term. For example,
greater NOx reduction could have been proposed. California
already has a 5.1 g/BHP-hr NOx standard for HDDEs. However,
California has no particulate standard for HDDEs and
particulate levels almost certainly average well above 0.60
g/BHP-hr. Since EPA also desired to establish near-term
particulate control, NOx controls were proposed only to the
point where they did not unduly impact potential near-term
particulate control levels.
Though always intertwined, the issues of feasibility and
leadtime are more separate here than in many other cases, due
to the fact that the 6.0/0.60 standards were proposed to take
effect in a very short period of time, just over two years from
the date of proposal. Thus, those issues related primarily to
feasibility will be discussed first followed by those concerned
primarily with leadtime.
a. Feasibility
• Overall, the 6.0/0.60 standards were fairly well received
by manufacturers. A number of manufacturers indicated that
they were feasible for either 1987 or 1988. Most
manufacturers, however, took issue with the details of EPA's
ana-lysis in support of the standards. Thus, these details need
to be addressed, as well as overall comments with respect to
feasibility and leadtime.
-------
2-35
These details fall into three basic categories. The first
is the identification of the design goal, or target, associated
with the two standards. The second deals with the projected
effectiveness of control technology and the ability to reach
the design targets. The third deals with the effect of these
technologies on BSFC, or fuel economy.
i. Design Targets
Design targets are a function of: 1) the emission
standard, 2) the DF applicable over the full useful life, and
3) emission measurement variability. The model, or equation,
used to determine a low-mileage target based on these
parameters is well known and accepted. The only issue relating
to the model itself is the assumption that the emission
-variability of an engine is known sufficiently well to allow
use of the z-statistic as an indication of the statistical
effect of this variability as opposed to the K-statistic.
Therefore, differences in estimated design targets arise due to
the use of different input DFs and emission variabilities or
the use of the K-statistic rather than the z-statistic.
A substantial amount of comment was provided on the
development of the target levels necessary for compliance with
the proposed 1987 standards. Six commenters provided numerical
value comments on the low mileage emission target levels
necessary for compliance with the proposed NOx and particulate
standards of 6.0 g/BHP-hr and 0.60 g/BHP-hr, respectively. The
target levels provided by the commenters are shown below:
Low mileage Target Level (g/BHP-hr)
Commenter
NOx
Particulate
Draft RIA
Cummins
EMA
Ford
GM
International
Harvester
Mack
5.1-5.5
5.25
4 . 94-light heavy
4.90-medium heavy
4.84-heavy heavy
4.5 to 4.9**
4.88-light heavy
4.84-medium heavy
5.45 for
product ion
variability
0.47-0.51
0.42 (assumes improved
knowledge)
0.35-light heavy
0.29-medium heavy
0.21-heavy heavy
0 . 32-0 . 47-med.ium heavy'
0.32 to 0.37**
0.36-light heavy
0.30-medium heavy
0.47 (0.05% sulfur in
fuel)
* «
Depending on assumptions on DFs and variability.
Target levels nay be increased as more knowledge is gained.
-------
2-36
The low mileage NOx target level was estimated in the
Draft RIA to be 5.1 to 5.5 g/BHP-hr, based on a coefficient of
variation (COV) for NOx and particulate of 10 and 10-15 percent
respectively, and full-life DFs of 0-0.48 g/BHP-hr NOx and
0.04-0.06 g/BHP-hr particulate and use of the z-statistic.
Neither Cummins nor Mack provided details on the methodology
used in developing their target level estimates, so their
estimates cannot be technically evaluated. However, both
estimates are inside the range identif-ied by the Draft RIA, at
least for NOx, so their estimates of the NOx DFs and COVs must
be close to those of the Draft RIA.
The EMA, GM and IHC estimates were based on use of the
K-statistic to account for emission variability, which assumes
that the standard deviation of NOx emissions for a given engine
family is unknown. As discussed in the Draft RIA, the more
appropriate statistic is the z-statistic, since fairly accurate
estimates of the standard deviation will be available prior to
production decisions for 1988. For example, as evidenced by
Cummins' comments, manufacturers are already testing their
production audit engines for particulate. No new information
was received which justified changing this conclusion.
The DFs used by EMA and IHC were derived from in-use
engine data from two sources: 1) an EEA study (performed for
EPA) of vehicle testing performed at SwRI and 2) engine testing
from the. joint EPA-EMA in-use test program. A linear
regression was performed on the after-maintenance data (or
as-received if maintenance was not performed) from these two
programs vs. mileage to derive DFs for NOx and particulate.
The resulting NOx DFs were not far from those estimated in the
"Draft RIA, but the particulate DFs were substantially larger.
Generally, such regressions are performed to derive
estimates of average in-use emissions. This was the purpose of
the EEA study sponsored by EPA. Included in the results of
such a study is an estimate of how fleet-average emissions
change with mileage (i.e., an in-use average DF). However,
unless the engines or vehicles tested meet the criteria for
inclusion in a recall action, the resulting DFs are not
appropriate for use in a design target analysis. -
An analysis of the engines included in these two programs
shows their condition to be far from satisfactory for recall
evaluation. Many were tampered and restorative maintenance was
performed on only 13 of 48 engines. Thus, the resulting DFs
essentially represent in-use DFs and not those •.:"
well-maintained engines and should not be used here.
-------
2-37
The GM DF for NOx was estimated from a subset of the data
referenced by EMA and IHC. However, even given this fact, it
fell into the range of the Draft RIA. GM's DF for particulate
was simply estimated to be 0.15 g/BHP-hr. This is well outside
the range used in the Draft RIA, but cannot be evaluated since
its basis is not known.
The Draft RIA NOx DFs were based on 1984 half-life data;
doubled to represent full-life DFs. Full-life 1985 data are
now available and are shown in Table 2-10. Only the
manufacturer-developed DFs are shown, since they are based on
actual durability testing. Assigned DFs are provided by EPA at
the manufacturers choice, but these are worst-case estimates to
encourage actual durability testing. Overall, half of the
developed DFs are zero and only three are significantly more
than the upper estimate used in the Draft RIA (0.48 g/BHP-hr).
The average developed DF in each subclass is 0.0 (LHDDE) , 0.1
(MHDDE) and 0.32 g/BHP-hr (HHDDE). Thus, the range of the
Draft RIA appears somewhat conservative for LHDDEs and MHDDEs.
Since quite a few HHDDE DFs are quite near 0.48 g/BHP-hr, the
Draft RIA upper limit appears quite appropriate for these
engines. It should be noted that manufacturers currently have
little pressure to reduce NOx DFs since the 10.7 g/BHP-hr
standard is well above low-mileage emission levels. Thus,
current DFs, particularly the largest, could very well
represent conservative estimates of future DFs when they become
a factor with respect to compliance.
Lacking data, the Draft RIA assumed the DF for particulate
emissions would be similar to the NOx or HC deterioration
factors when expressed as a percentage of the emission level.
Commenters contended that normal wear in such components as the
^11 o 1 -in^^^^Ton r\i i mm ^ *• e r^rNnf-yole in-i^^t*/^r*c a n H n T c t- r% n T-Tnnc
^ «» >« * .».*J >»*•*..***.» £-«••••£-, ***** WV..WKW~W, «.».JWW»W — w «»..*_ ^- _ w w .w . . _ _ . . -3 v
would be expected to cause an increase in HC and/or particulate
emissions while causing a decrease in NOx emissions. Given the
fact that most NOx DFs are zero or negative and this would not
be expected for particulate, the use of HC DFs as a surrogate
is probably more appropriate. Referring to Table 2-11, for
LHDDES, the ratio of the mean deterioration factor to the mean
low mileage emission level was found to be 0.14. Corresponding
ratios for MHDDEs and HHDDEs were found to be 0.05 and 0.06,
respectively. Under a particulate standard of 0.60 g/BHP-hr,
the low mileage level will be roughly 0.5 g/BHP-hr and the
preceding ratios developed from actual HC deterioration factors
would correspond to particulate deterioration f-actors of 0.07,
0.025, and 0.03 g/BHP-hr for light, medium and heavy HDDEs,
respectively. These values bracket very closely the DF range
(.04 to .06 g/8HP-hr) developed in the Draft RIA. Thus, tnis
range continues to appear appropriate.
-------
2-38
Table 2-LO
1985 Federal Full-life Deterioration Factors
LHDDE DF (g/BHP-hr)CI]
GM 0.0
IHC 0.0
MHDDE
GM 0.0, 0.0
Caterpillar 0.02
IHC 0.0, 0.0, 0.61
HHDDE
GM 0.65, 1.14
Caterpillar - 0.0, 0.0, 0.47
Cummins 0.07, 0.39, 0.46, 0.46
Mack 0.00, 0.00, 0.00, 0.37
Volvo White 0.50
Only manufacturer-developed DFs are shown. Assigned DFs
are essentially worst-case DFs and are not necessarily
indicative of an engines actual DF.
-------
2-39
Table 2-11
1985 Model Year HDDE HC Emission Levels and
•Deterioration Factors Developed by Manufacturers
Manufacturer HC DF HC Low-Mileage Emissions
LHDDE
General Motors 0.15 0.53,0.46
International Harvester 0.00 0,79
MHDDE ^ '
Caterpillar 0.06 • 0.62
General Motors 0.05,0.00 0.58,0.84
International Harvester 0.00,0.08,0.03 0.70,0.85,1.32
t .
HHDDE
•—- -Caterpillar , . " . . 0.00,0.21,0.01 0.19,0.36,0.32
Cummins - 0.00,0.00,0.00,0.02 "0.46,0.62,0.92,0.52
General Motors 0.00,0.00 0.48,0.54
Mac* 0.19,0.00,0.00,0.00 0.90,0.69,0.74,0.54
Volvo White 0.10 0.81,1.15
-------
2-40
With respect to the last pertinent factor, emissions
variability, EMA, IHC, and GM all used estimates which included
lab-to-lab variability. This would be appropriate in an
analysis focused on pre-production certification requirements,
if EPA were to perform confirmatory tests at its own lab.
However, the focus here is SEA, because its requirements are
statistically more stringent than those of certification. SEAs
are performed at manufacturers' own facilities. Thus, any
differences between a manufacturers' own labs are well known
and characterized. Thus, inclusion of lab-to-lab variability,
particularly insofar as these estimates were based on the
variability among seven independent test facilities, is .not
appropriate here. When this is taken into account, the
estimates of EMA, IHC, and GM would be very similar to those of
the Draft RIA.
Overall, then, the inputs parameters estimated in the
Draft RIA still appear appropriate. Thus, the design targets
.remain unchanged at 5.1-5.5 g/BHP-hr NOx and 0.47-0.51 g/BHP-hr
particulate. However, the facb that most manufacturers'
estimated design targets were well below these levels should be
kept in mind below as control technology effectiveness is
discussed. An unrealistically low design target overestimates
the degree of control necessary to achieve a standard.
Therefore, either the necessary application of technologies is
overestimated, or a standard is termed infeasible when it is
not.
ii. Control Technology Assessment
The analysis of HDDE control technology is inherently
difficult, because each manufacturers' engines are designed
somewhat differently and have varying technical capabilities.
Differences between the generic HDDE subclasses compounds this
task. Thus, engine-specific analyses are not possible due to
the complexity of the task. However, even if such an attempt
were possible, the necessary data are not available in most
cases. Thus, the analysis in the Draft RIA and that performed
here must address generic control techniques and reduction
capabilities, while at the same time considering differences
between engine designs insofar as possible.
Another factor adding to the complexity of the task is the
rapid change in technology currently affecting HDDEs. • New
technologies, such as enhanced aftercooling, variable injection
timing, electronic engine controls (EEC), higher-pressure
injection and higher efficiency, faster response turbochargers
are all being introduced to some degree to improve BSFC,
regardless of emission levels. However, many of these
technologies also directly effect NOx and particulate and are
-------
2-41
among those considered below as potential control
technologies. A problem is tha€ all of these can be optimized
for BSFC or emissions and interact in a complex way. Thus, it
is also difficult to determine a pre-control baseline. The
result is that future technology must be estimated both with
and without these standards and data from engines encompassing
a representative sample of these technologies must be relied
upon to estimate overall control effectiveness. While
important here, these factors are even more dominant in the
analysis of the 4.0/0.40 standards to follow.
Unfortunately, little data were received in comments on
the 1987 standards which quantified the effect of the various
control techniques projected to be both available and effective
in achieving these standards in the Draft RIA. Most commenters
simply stated whether or not the 6.0/0.60 standards were
feasible and, if so, when. Some also presented their
qualitative judgment of EPA's feasibility analysis. A few
(e.g., GM) presented charts of NOx/particulate and NOx/BSFC
curves for each of their engines. However, without test data
and descriptions, these also cannot be properly evaluated.
Manufacturers' comments pertaining to overall feasibility will
be summarized first, followed by general comments, pertaining
to the Draft RIA analysis. These comments will then be
analyzed using what data were supplied, as well as those
included in the Draft RIA. - . ...
Daimler-Benz stated, without qualification, that their
MHDDEs could achieve compliance with the proposed standards in
*:1987, as did Volvo White with respect to 1988. Ford also
".stated that" compliance with the proposed 1987 standards was
• achievable, but indicated this conclusion was based on
projections chat both DFs and emission variability would be
relatively low (which they expected and which appeared
reasonable given the analysis presented above). GM indicated
that its medium- and heavy-HDDEs could also comply in 1987, but
with some fuel economy penalty (which is addressed below). in
the case of its light heavy-duty engine, GM indicated that
compliance with both the proposed particulate and NOx standards
was not achievable simultaneously.
Comments by the other manufacturers as well as by EMA did
not include direct statements on either an anticipated ability
to comply nor an anticipated inability to comply. However, the
comments did include discussions of the -technologies required
for compliance and the time required for implementation. Thus,
it is reasonable to infer that compliance with the proposed
standards was considered to be technically achievable by these
other manufacturers, as well. Cummins and Mack did mentis
implementation years of 1989 and 1990 respectively, for j-.
least 'the NOx standard. However, leadtime will be considered
further below.
-------
2-42
Overall, the only manufacturer to absolutely question the
feasibility of the 6.0/0.60 standards was GM for its LHDDEs.
On the surface, this is rather surprising since data generated
in EPA's Ann Arbor Lab on a low-NOx version of this engine
(referenced in the Draft RIA) showed it to have the lowest
combination of NOx and particulate emissions of any engine (3.0
NOx and 0.46 particulate, g/BHP-hr). Also, prototype data
submitted by GM after the original proposal (then confidential,
but recently made public in an EPA-sponsored study[6]> show
emissions to be 4.1 NOx/0.46 particulate and 2.8 NOx/0.52
particulate at two calibrations (all in g/BHP-hr). While the
levels of 1984 production engines are somewhat higher (4.2
NOx/0.66 particulate and 3.6 NOx/0.62 particulate,
respectively), these levels are still low relative to those of
the other engines and no incentive existed in 1984 to keep
either NOx or particulate as low as the prototype levels.
GM did not refer to any of these data, but did present a
NOx/particulate trade-off curve for this engine. The curve is
slightly below the 1984 production data, but well below
prototype curve. No explanation is given concerning the
prototype/production difference. Also, GM's estimated design
target for the particulate standard is 0.32-0.36 g/BHP-hr,
which is below even the prototype data and may explain GM's
conclusion. It should not be necessary, based on EPA
estimates, to design an engine below a design target of
0.47-0.51 g/BHP-hr particulate, as discussed above.
Consequently, this engine must be considered capable of
complying with the proposed standards.
Moving on to comments on the Draft RIA analysis, a number
of manufacturers (Caterpillar, in particular) indicated that
some of the analyses were rather simplistic and not realistic.
For example, Caterpillar took issue with EPA's statement that
California's 5.1 g/BHP-hr NOx standard could easily be met with
simple injection timing retard. Caterpillar also disagreed
with EPA's implication that transient particulate emissions can
be reduced to steady-state levels, through improved transient
fuel rate control. • •
With respect to the first statement, Caterpillar took the
statement more literally than intended. The primary point
being made was that, with respect to techniques designed
primarily to control NOx control techniques, injection timing
was sufficient (i.e., no other NOx control techniques were
required) and the point was not that absolutely no other
changes (e.g., recalibrations) would be required. Caterpillar
lists a number of changes made to its California engines in
addition to injection timing retard, such as power de-rate,
-------
2-43
turbocharger modifications, and fuel governing modifications.
These are reasonable recalibrations whenever a basic engine
parameter, such as injection timing, is changed. However, they
in themselves are not necessarily NOx control techniques,
though their cost must be considered.
With respect to the second statement, Caterpillar's
judgment is based on a belief that advances in turbocharger
design have already achieved most of what is to be gained " in
improved transient response. Further, they claimed that
over-fueling is necessary to accelerate an engine. Again, the
point being made in the Draft RIA was not that the entire
transient/steady-state difference could be eliminated, but that
improvements were possible and the current
transient/steady-state difference was an indication of this
potential. Given the work known to be underway by -both
turbocharger manufacturers and other HDDE manufacturers—
-evidenced by the numerous technical papers in the area even
though most of what is being done is proprietary—it does not
appear reasonable to conclude that turbocharger response cannot
be measurably improved. Also, the potential capability of
electronics to precisely limit fuel delivery to minimize any
particulate control/performance could also be substantial.
Whether such improvements can- be achieved by the 1987-1988
timeframe fleet-wide is another issue.
Each manufacturer also identified, in varying degrees' of
detail, the technologies which it expected to use on one or
-more of its engines to achieve compliance with the proposed
6.0/0.60 standards. The technologies identified were as
follows: 1) application of turbocharging, 2) turbocharger
modifications, such as improved efficiency and transient
response, 3) addition o£ a£tercooling to turbcchargsd engines,
-4) enhanced aftercooling), 5) injection timing retard, 6)
addition of variable injection timing, 7) increased fuel
injection pressure, 8) fuel injector modifications, and 9)
modifications to the combustion chamber and air swirl rate.
Manufacturers also indicated that they anticipated an ongoing
introduction of electronic controls focused mainly on the
minimization of fuel economy penalties.
These technologies are basically the same as those
projected in the Draft RIA for both the 1987 and 1990
standards. While some use of the technologies associated witn
the latter standard was anticipated in. 1987, manufacturers
appear to be utilizing a greater number of combinations of
technologies at the 6.0/0.60 level than had been projected in
the Draft RIA, possibly because of fuel economy concerns.
-------
2-44
Telephone communications with manufacturers concerning
their 1985 model year California engines showed that
combinations of the technologies listed above are in use on
these engines. However, since the half-life California NOx
standard is essentially equivalent to a 5.1-5.35 g/BHP-hr
full-life standard, it is 0.65-0.9 g/BHP-hr more stringent than
the proposed Federal standard and not all of these technology
modification/additives (at least these which are NOx related)
should be required to comply with the 6.0/0.60 standards.
With respect to the NOx standard, EPA acknowledges that
sole reliance on injection timing retard to achieve compliance
in the 6.0 g/BHP-hr NOx standard could result in significant
fuel penalties. Thus, to minimize fuel penalties manufacturers
may elect to increase the use of aftercooling and variable
injection timing. However, the 'use of enhanced aftercooling,
particularly air-to-air units, appears more appropriate at NOx
levels more stringent than 6.0 g/BHP-hr, it should not be
necessary at 6.0 g/BHP-hr NOx. If air-to-air aftercooling were
applied, it would be to reduce BSFC and should not be included
as a cost of the 6.0 g/BHP-hr standard.
With respect to particulate emissions, some additional use
of turbocharging was projected in the Draft RIA, particularly
with respect to Caterpillar's 3208 engines. This was confirmed
by Caterpillar in their comments. Also, the Draft RIA
identified the general need for modifications to existing
engine components, but none involving additional hardware.
These components include modified injectors and combustion
chambers, improved fuel governing during transients, and
moderate increases in injection pressure, all of which are
described in more detail in the Draft RIA.
Due to the difficulties mentioned above, such as
heterogeneous engine designs, lack of engine-specific data and
rapidly changing technology to reduce BSFC, specific estimates
of the technological changes necessary for each engine cannot
be made. However, most of the changes described above
primarily involve research, development and tooling. The
revised components should inherently be no more expensive in
the long-run than the original components. Thus, the cost of
these standards may not depend strongly on the number of
changes made, but rather on the need to perform the necessary
research and development to determine which changes actually
.need to be made. For the most part, much of this research has
been ongoing already or performed.
-------
2-45
iii. HC and Fuel Economy Effects
V
No technically supportable comments were received
indicating the 1988 standards would significantly increase HC
emissions. However, comments pertaining to the fuel economy
effects of the proposed standards were provided by most
commenters. The estimated fuel economy penalty anticipated by
each comrnenter are shown below together with the basis for the
estimate, when one was provided.
Commenter Fuel Economy Penalty
Caterpillar 3-12 percent -(1985 Federal/
California difference)
Cummins 1-3.5 percent -
GM .. _ • 3-5 percent for MHDDE
4-6 percent for HHDDE '"'
2 percent for new design HHDDE
International Harvester 4-8 percent from NAS study
5.4-7.2 percent MHDDE
7.7-8.3 percent HHDE (1985
Federal/California difference)
Mack 6 percent (1985 Federal/ California
difference is 4.7-12.5 percent)
Daimler-Benz • ... No significant loss in fuel economy
v ' Many of the manufacturers' projections on reduced
efficiency of fuel utilization were based on differences
between 1985 model year Federal and California engines. This
is not an appropriate comparison since the California standard
is 5.1, not 6.0 g/BHP-hr and the California engines are 1985
models, not 1988. The California standard is a full-life
standard, but does not include an assembly line test program.
This allows a somewhat smaller safety margin, since SEA is
generally considered to be statistically more stringent than
certification. However, this difference should be only a small
part of the 0.5 g/BHP-hr margin attributable to SEA. Thus, the
California program can be considered to be essentially on par
with the Federal program at equal standard levels. As the 5.1
g/BHP-hr NOx standard is much closer to - 5.0 rather than 6.0
g/BHP-hr, the California data are more useful in assessing the
fuel economy effect of the 5.0 g/BHP-hr standard than this o.ne.
IHC made an additional comment that the NAS study was not
based on "old"- technology, as concluded in the Draft RIA, but
on advanced engine technology, at least insofar as the data IHC
supplied. Whether or not this is true for other manufacturers'
-------
2-t6
data cannot be determined, since no other comments were
received on this point. However, even if some of the data were
from advanced technology engines, the absolute NOx and BSFC
levels of the NAS curve and the resulting trade-off make it
apparent that any advanced technology was applied to optimize
BSFC and power and not NOx control. This approach is not
consistent with the approach taken in estimating the economic
impact of this rule (see Chapter 3), where the cost of advanced
technology is being charged co NOx control. Given this, any
fuel penalty should be determined from the BSFC of the engine
without the advanced technology, not with it. Also, the nature
of the NAS study made it impossible to cite the specific data
used to derive their NOx/BSFC curves. Therefore, it cannot be
determined how much optimization of BSFC occurred at low NOx
levels. Thus, the NAS BSFC/NOx curve still appears to
overestimate the effect of NOx control.
There is another reason why some of the manufacturers'
estimates shown above may overestimate the fuel economy penalty
of the 6.0 g/BHP-hr standard. That is the fact that at least
GM and IHC used a low-mileage NOx design target 0.6 g/BHP-hr
lower than that necessary. Use of a 5.1-5.5 g/BHP-hr target
would result in lower estimated fuel penalties. This appears
to be confirmed by Cummins estimate. Cummins' estimated a NOx
design target with the above range and also projected the
lowest fuel economy penalty of any manufacturer, except
Daimler-Benz.
While 1985 California BSFC penalties cannot be directly
applied to the 6.0 g/BHP-hr standard, they can be used
indirectly to confirm the 0-2 percent estimate of the NPRM.
These differences between the California and Federal situation
need to be considered. One, an additional three years of
leadtime will be available allowing additional control system
optimization . Two, the low-mileage targets will be 0.9
g/BHP-hr NOx higher so less NOx control will be required,
lowering any fuel penalty. Three, the technologies being used
in California are primarily quick fixes, requiring low initial
capital investment (research, development, soiling). Given the
longer leadtime available and the fact that nationwide sales
will be effected by high BSFC, and not just California sales,
much more comprehensive research and development, resulting in
optimized control approaches and lower BSFC penalties, should
result even with today's technology. Thus, the upper end of
the current California penalties must be considered extreme
under these conditions. The lower end of the penalties, 3-4
percent, should also be able to be lowered, given the
additional leadtime and added return for the same investment
(i.e., Federal vs. California sales). Thus on average in the
short term, a maximum penalty of 2 percent may result from the
6.0/0.60 standards. The possibility of no penalty also exists
given Daimler-Benz' comment. Thus, in the short term, the
average fuel economy penalty should be 0-2 percent.
-------
2-47
In the long run, beyond 1988, one would expect the penalty
to disappear altogether. This is because the advanced
technology projected in the 1991 timeframe should improve fuel
economy such that any short-term penalty will be eliminated by
the early 1990's. Also, general improvements in BSFC will
lower fuel consumption over the emissions test cycle and, other
things being equal, NOx and particulate will decrease as well.
It was estimated in the Draft RIA that BSFC would decrease
roughly 1.5 percent per year in this timeframe based on
comments from MVMA and EMA to the MOBILES development process
confirm this figure. Thus, three additional years should
provide a 4-5 percent reduction in NOx emissions simply due to
BSFC improvements.
b. Leadtime
The issue of the amount of necessary leadtime associated
with the 6.0/0.60 standards received a substantial comment.
This analysis will begin by describing the steps necessary in
developing engines and vehicles to meet emission standards,
along with the time associated with each step. The comments to
the 1987 implementation date will then be summarized, followed
by an analysis of those comments.
All work necessary for emissions compliance by the engines
does not have to be completed prior to initiation of design
work for engine integration into the vehicles, but sufficient
progress has to be made in engine development so as to clearly
define the engine envelope (overall spatial requirements of the
engine, including aftercooling). The primary tasks involved in
the successful development and marketing of engines complying
with the 6.0/0.60 g/BHP-hr standards in vehicles are shown in
Table 2-12.
The total leadtime requirement for engine development is
the sum of tasks A through H less task C, or 31-38 months.
Since the standards are applicable to all HDDEs, the greater of
the two time requirements for durability data development was
used. In the case of vehicle development, the leadtme
required is the sum of tasks A, B, C, I, J, and K, or a total
of 28-36 months.
Starting with a March 15, 1985 date for publication of the
final rule, the time available for implementation of new
standards by the 1987 model year. (January 1, '1987) would be
approximately 21 months. The time available for implerrentation
of new standards by the 1988 model year (January 1, 1988) would
be approximately 33 months. Since the time available for
implementation by the 1987 model year is significantly less
-------
2-48
B
D.
E.
G.
H,
Table 2-12
Leadtime Projection - 6.0/0.6 Standards
Task
Identify, by engine, technologies probably
required for compliance, develop initial
designs
Procure initial design hardware, build and
test initial design test engines
Develop engine envelope requirements for
vehicle builders
Develop second level engine designs
Procure hardware, build engines with "
alternative calibrations and develop
emission data and fuel economy
characteristics by calibration to define
durability engine calibrations
Develop emission durability data
Develop data from emission data engine
Coordinate emissions certification
compliance with EPA
Develop overall vehicle design considering
the effects of all engines offered in each
vehicle
Procure new dies for the manufacturer of
redesigned vehicle components
Confirm mechanical durability of redesigned
vehicle components
Time Required
3-4 months
4-6 months
1-2 months
2-3 months
8-9 months
4-5 months
for light-
heavy
11-12
months for
heavy-heavy
1 month
2-3 months
8-10 months
7-9 months*
5 months**'
**
Reference 2.
60,000 miles at 40 mph average speed, two effective 7-hour
shifts per day an-1 six days per weeX.
-------
2-49
than even the minimum time estimated above, implementation in
the 1987 model year does not.ap'pear feasible for most engines.
The time available prior to the 1988 model year is within the
range of estimated time required. Also, the entire process can
be accelerated in those extreme cases where more time is
necessary. Thus, 1987 should be ruled out on the basis of
inadequate leadtime for industry-wide compliance; however, the
1988 model year appears feasible.
Moving to the comments, one commenter, Daimler-Benz,
stated that their engines could be brought into compliance with
the proposed standards for the 1987 model year. GM stated that
all but their one LHDDE could comply in 1987. However, other
.commenters indicated that a greater amount of time was
required, usually one year, but occasionally more.
Specifically, Ford and Volvo White acknowledged that 1988 was
feasible while Caterpillar indicated that 1988 was the earliest
date feasible. International Harvester estimated that 39
months starting from the date that the engine configuration is
finalized would be necessary to allow integration of the
reconfigured engine envelope into the vehicle, to accommodate
changes in engine cooling requirements, the addition of
air-to-air aftercooling, the addition of electronics and
compliance with noise and safety standards.
Cummins and Mack requested that the standards be delayed
until 1989 and 1990, respectively. Both cited the statutory
mandate of 4 years leadtime, but also referred to technical
difficulties. Cummins indicated generally that anything less
than the statutory leadtime would require them to accelerate
development of their planned engine modifications to a degree
which would seriously affect the durability, reliability and
fuel efficiency o£ their engines. Mack considered the 1990
date necessary because essentially all Mack engines had to be
redeveloped and personnel limitations precluded earlier
completion.
The two projections (by Daimler-Benz and GM) of 1987 as
the feasible year of introduction indicates the ability to
compress the schedule described in Table 2-11 above. It may
also indicate that manufacturers are starting from different
points (i.e., levels of current emissions).
Without emission data or specific leadtime estimates, it
is impossible to evaluate the IHC, Cummins and Mack leadtime
estimates, which are the only ones requesting time beyond
1988. (Cummins' and Mack's legal agreements are addressed in
the Preamble to the FRM.) Generally speaking, the types of
changes being referred to by IHC should not be necessary co
comply with the 6.0/0.60 standards. They may be desirable in
the long-run to mprove BSFC, but they are not driven by the
6.0/0.60 standards.
-------
2-50
Given the above leadtime analysis, the infeasibility of
1987 as an effective, date, the general support for 1988 as a
feasible year of implementation, and no clear, supported
arguments by any manufacturers against it, 1988 is determined
to be the year the 6.0/.60 standards should be implemented.
2. Mid-Term NOx and Non-Trap Particulate Standards
A mid-term (1990) NOx standard of 4.0 g/BHP-hr was
proposed for all HDDEs. A 1990 particulate standard of 0.40
g/BHP-hr was also tentatively identified as the non-trap
technological limit, and proposed as a possible standard for
non-urban (linB-haul) HDDEs. In identifying these levels, the
same approach was used as that described above concerning
development of the near-term standards. The goal was to obtain
both NOx and particulate emissions, but NOx emission control
was balanced so as not to severely impact the ability to
control particulate emissions.
The difficulties in performing an analysis such as this,
which were described with respect to the 6.0/0.60 g/BHP-hr
standards above, apply even more here. While the heterogeneity
of engine designs is the same, technology is changing even more
dramatically in this later -timeframe and the interaction
between control techniques is even stronger. Also, even less
data exist than was available for 1987 technology, to base
feasibility judgments on.
Again, as with the 1987 standards, there are two basic
issues: technical feasibility and leadtime. However, here the
issues associated with leadtime are much less significant,
because the implementation date is sufficiently distant to
allow significant research and development application. The
long leadtime available should provide manufacturers with
adequate opportunity to overcome problems and undesirable
effects associated with additional NOx control. Also, the
proposed 1987 standards require delay until 1988 and the Act
requires a three-year interval for NOx emission standards;
thus, the mid-term NOx standards cannot take effect until
1991. This analysis will presume simultaneous implementation
of both NOx and particulate standards since that will maximize
the manufacturers' ability to design engines that can meet both
standards.
a. Technical Feasibility
The analysis of these technical comments will follow that
for the 6.0/0.60 g/BHP-hr standards. The one exception is that
no reanalysis of design targets will be performed here. No new
information is applicable that was not already discussed with
-------
2-51
respect to the 6.0/0.60 standards. That analysis confirmed the
Draft RIA's targets. Thus, the Draft RIA NOx target of 3.2-3.6
g/BHP-hr will be used below. A target associated with the 0.40
g/BHP-hr particulate standard was not explicitly determined in
the Draft RIA, but it would be about 0.30-0.33 g/BHP-hr. The
only point to keep in mind is that the targets manufacturers
used in assessing the feasibility of the 4.0/0.40 g/BHP-hr
standards are, for the most part, significantly lower than
those deemed necessary here. Therefore, their statements may
exaggerate the feasibility efforts of various standard levels.
Every manufacturer stated that the 1990 4.0 g/BHP-hr NOx
and 0.40 g/BHP-hr particulate standards were 'not
technologically achievable with any combination of known or
anticipated technologies. While the analysis of the Draft RIA
for both the 4.0 g/BHP-hr NOx and 0.40 g/BHP-hr particulate
standard identified a number of potential control technologies
in each case and used available data to roughly estimate the
potential control efficiency of each technique, no commenters
presented parametric studies of any of these technologies which
would better demonstrate their potential effectiveness. Many
general comments argued against the efficiencies estimated in
the Draft RIA, based on technical grounds, but without the
depth of analysis necessary to fully support the point being
made and negate the point of the Draft RIA. Thus, while a
degree of doubt has been thrown on the estimates of the Draft
RIA, insufficient data are available to go through the NPRM
analysis point by point and reestimate the effect of each
technology.
However, some confidential data were made available
indicating the combined effect of a number of these techniques
(e.g., increased injection pressure and enhanced affcercooling).
as well as the projections of the levels feasible in this
timeframe. while such data cannot be used to directly
determine the fullest potential of one or more technologies,
they do represent the most quantitative set of estimates
available. Some degree of evaluation can be applied using the
estimates of the Draft RIA. To facilitate their use here,
these data have been combined into a single figure (Figure 2-2)
which shows both current levels of NOx and particulate
emissions and the manufacturers anticipated achievable levels.
Superimposed on the best achievable emission control
projections are the range of low-mileage targets as previously
developed for the 6.0/0.60 standards (points At and A,) and
the midpoint of the low-mileage targets for the engine-out
standards of 4.0/0.40 (point E) . A low-mileage target level ::"
4.2-4.6 g/BHP-hr for a NOx standard of 5.0 g/BHP-hr (vert.-31
lines B) was also developed by the same procedure and i!-e
midpoint shown for comparison purposes. Particulate leveis L
-------
0.9 .
0.8 .
0.7 -
1'arLiculdLe
EMISSIONS
g/BIBMu:
0.5 J
0.4 .
0.3 -
q.2 -
0.1 -
figure 2-2 ,
Ileavy-Puty Engine NQx and Particulate Enqina-Oub Characteristics
E
n
CCurrent Engines
Best Teclinologies
V
2 J
7 8
10 11
NQx Qiiissions g/DIIP-hr
-------
2-53
and D correspond to the intercepts of lines C with the upper
and lower bounds of the projected range of emissions using best
available technology combinations as estimated by the
manufacturers.
The first observation to make about the best technology
estimates is that the upper limit barely passes through the
design targets for the 6.0/0.60 standards. Thus, it represents
fairly near-term "best technology." Second, the lower limit
does not even approach the targets for the 4.0/0.40 'standards.
Given this, it is reasonable to evaluate the lower limit of the
best technology curve against the projections made in the Draft
RIA.
First, with respect to NOx, the Draft RIA analysis of 4.0
g/BHP-hr NOx relied upon large NOx reductions at constant BSFC
for separate circuit and air-to-air aftercooling. This
estimate was based on data from one GM/DDA engine and involved
estimating NOx reductions beyond that evidenced by the data
based on an estimated BSFC/NOx tradeoff for timing retard.
Thus, no actual NOx data below 5.0 g/BHP-hr were available.
Complicating matters was the absence of any particulate data in
the study. While it can be assumed that these were not above
0.80 g/BHP-hr since BSFC was improving, particulate may have
been well above 0.60 g/BHP-hr. Thus, these data may not be
inconsistent with the curve in Figure 2-2, the problem may be a
lack of particulate control, not NOx control.
• Based on testing performed on the same DDA engine, the
Draft RIA estimated that electronic engine controls (EEC) also
had the potential for large NOx reductions at constant BSFC.
Again, however, the estimate involved assuming a BSFC/NOx
tradeoff curve for the engine and using timing rstsrd sgsir.st s
BSFC improvemenc to estimate NOx emissions at constant BSFC. A
6.0 g/BHP-hr NOx level was the lowest actual data point in this
analysis.
Many manufacturers stated that this analysis overstates
the benefit of EEC. Some argued that the NOx benefit of EEC
depends on the final NOx level (i.e., its benefit is large at
6.0 g/BHP-hr NOx and negligible at 4.0 g/BHP-hr). Others
argued that the benefits of enhanced aftercooling and EEC were
mutually exclusive, due to the fact that combustion efficiency
limits the use of either technology and determining when both
BSFC and particulate emissions begin to increase dramatically.
Without data or sophisticated combustion analysis, it is
impossible to prove or disprove these comments. However, these
data could be consistent with that in Figure 2-2, given that
particulate emissions are unknown.
-------
2-54
With respect to particulate control, the Draft RIA
analysis was much more general and based on less data than that
for NOx, as 0.40 g/BHP-hr was not the primary, proposed
standard. The most feasible technique for the 1990 timefrarre
was high-pressure fuel infection. The only data available were
steady-state emissions on one engine. The other technologies
were 1) injection rate modification and/or modulation which
would require significant advances in injector technology, 2)
ceramics, which are not progressing as fast as some had
projected a year ago, and 3) conversion to methanol fuel, which
though technically feasible requires the establishment of a
fuel distribution system (except for buses, as discussed
below). Otherwise, small improvements could .be expected from
general BSFC improvements, and for additional optimization of
injectors and combustion chambers. Thus, little data exist
with which to refute the data in Figure 2-2 and it must be
taken as the best estimate of technology in the 1991 timeframe
at this time.
Given this, a 5.0 g/BHP-hr NOx standard for 1991 appears
most reasonable instead of the proposed 4.0 g/BHP-hr standard.
This is principally because of the adverse tradeoff between NOx
and particulate which appears likely. While a 4.5 g/BHP-hr NOx,
level may be potentially achievable, particulate emissions
appear to begin increasing at a distinctly higher rate below
5.0 g/BHP-hr NOx. Also, below 5 g/BHP-hr NOx, particulate
emissions could in some cases increase well above 0.6 g/BHP-hr,
making trap application very difficult. This would be
particularly true with respect to the 1991 0.1 g/BHP-hr
particulate standard for buses. The BSFC tradeoff would also
begin increasing dramatically here, as well.
With respect to particulate, the range of expected low
mileage engine-out particulate levels corresponding to the
target level required for a NOx standard of 5.0 g/BHP-hr would
be 0.42-0.54 g/BHP-hr (levels C and D in Figure 2-2). If it is
assumed that trap-based standards are implemented in 1991 and
1994 (i.e., pressure to control particulate continues through
the 1994 timeframe), then it 'is likely that progress will
continue to be made in reducing engine-out particulate levels
down to the lower best technology' curve, resulting in an
engine-out particulate level of 0.43 g/BHP-hr by 1994. If
stringent 1991 non-trap particulate standard were implemented
in 1991, it should also be able to reduce engine-out levels to
0.42 g/BHP-hr, but three years earlier. Thus, at best, a 1991
non-trap standard 0.50 g/BHP-hr appears achievable.
These levels are at least partially proposed by two
manufacturers. Cummins recommended (at the public hearing on
the NPRM) 1992 target standards of 4.5 g/BHP-hr NOx and 0.50
-------
2-55
g/BHP-hr particulate. Daimler-Benz recommended 1990 standards
of 5.1 g/BHP-hr NOx and 0.68 g/BHP-hr particulate. Other
conmenters either did not recommend any alternative standards
to 4.0/0.40 g/BHP-hr or recommended retaining the 6.0/0.60
g/BHP-hr standards indefinitely.
b. Effect on Fuel Economy
As all commenters stated that the 4.0 g/BHP-hr NOx
standard was infeasible, no estimates of its effect on fuel
economy were made. Nor were any comments received addressing
the effect of a 5.0 g/BHP-hr NOx standard. However, the latter
can be estimated from the 1985 California/Federal comparison
conducted above.
As in that section, there "are a number of differences
between the 1985 California 5.1 g/BHP-hr standard and the 1991
Federal standard of 5.0 g/BHP-hr. First, the Federal
low-mileage target is slightly more than 0.1- g/BHP-hr lower
than that in California. This would tend to slightly increase
the fuel economy penalty. Second, six years of leadtime exist
between the California and Federal situations to develop
improved technology. Third, any adverse BSFC effects would
affect Federal sales, which is roughly 10 times larger than
California's. Both the available leadtime and the potential
national sales impact, manufacturers would be expected to do
all that is possible to eliminate any BSFC effect, as opposed
to the California approach, which is more short-term, quick-fix
oriented.
Given that significant NOx control technologies such as
electronics, separate-circuit aftercooling, and air-to-air
aftercooling are not currently present at all in California,
and given their projected widespread use by 1991, it would
appear that the additional leadtime and potential national
impact would overwhelm the first. Overall, it would not appear
unreasonable to project the same long-term fuel economy impact
here as that projected in the NPRM for the 4.0 g/BHP-hr
standard, zero percent. However, due to uncertainty in this
analysis, and the projection that the BSFC/NOx curve begins to
turn sharply upward at approximately a 5.0 g/BHP-hr standard, a
long-term 0.5 percent fuel economy penalty may occur. In the
short-term, a slightly higher 1.0 percent penalty may be
experienced.
" 3. Particulate Traps
In response to the NPRM, the Agency received a large
number of comments directed towards its heavy-duty trap
feasibility analysis. As explained above in the synopsis of
-------
2-56
the NPRM analysis, the Agency concluded that traps were
feasible for 1990 model year HDDEs, extrapolated from the
status of light-duty trap technology and the design effort
necessary to adapt this technology to heavy-duty usage. The
comments were both supportive and critical of EPA's analyses
and conclusions. This part of the regulatory impact analysis
will respond to the comments, concentrating on the points of
the analysis with which the commenters disagreed. New
information that is pertinent to the heavy-duty trap
feasibility question will also be incorporated. The comments
will be addressed in the same format used in the previous
analysis: light-duty trap status; LD/HD differences;
heavy-duty trap status; and emission levels. .In addition, the
leadtime issue will also be addressed.
a. Light-Duty Trap Status
The status of light-duty trap oxidizers was generally not
addressed by the commenters. The notable exception was General
Motors. GM's position is that technology is still not
available to meet the promulgated 1987 light-duty vehicle and
light-duty truck standards. The extensive LOT testing (200
alternative fuels and fuel additives combined with over 150
trap materials in over 500 traps) conducted by GM has not
resulted in an identification of a LD trap that can be
committed to a production program. Thus, GM strongly objected
to EPA's conclusion that light-duty traps are technically
feasible for 1987 model year vehicles.
General Motors' comments notwithstanding/ the Agency's
position in the NPRM was borne out by Mercedes-Benz's
certification of its 3L turbodiesel, equipped with a trap to
meet the California Air Resources Board (CARS) 1985 model year
standards.[8] CARB's 1985 standard for California light-duty
diesel vehicles is 0.40 grams per mile (g/mi) particulates, to
be further reduced to 0.20 g/mi in 1986 and 0.08 g/mi in 1989.
In addition to its 1985 California LDDVs (which are also sold
in other western states), Mercedes-Benz plans to add traps to
all its U.S. sold 3L LDDVS in 1986, a year prior to the
promulgated 1987 0.20 g/mi standard.
Mercedes-Benz is not alone in certifying a trap-equipped
LDDV. Volkswagenwerk AG (VW) plans to install a trap on its
larger diesel LDV (Quantums) in California beginning in tne
1986 model year.[7] VW intends to equip - all federally
certified Quantums with trap-oxidizers the following year to
comply with the 1987 LDV particulate standards. The trap
applications of Mercedes and VW are proof that trap-oxidize..,
are a viable form of light-duty particulate emissions control.
-------
2-57
b. Light-Duty/Heavy-Duty.. Differences
The manufacturers which commented agreed with EPA's
analysis of the differences between light- and heavy-duty
applications that must be considered in the design of a
heavy-duty trap oxidizer. Comments from the manufacturers
restated these differences (engine size and load factor,
operating conditions and temperatures, durability, and ash
accumulation), adding very little to what was previously
reported in the draft analysis. The design efforts continue to
be directed towards a suitable regeneration system that can
handle the increased exhaust flow of the heavy-duty engine
environment and the generally lower exhaust temperatures of a
turbocharged engine. The commenters believe that the greatest
design challenge is the required durability of a heavy-duty
trap as opposed to a light-duty trap.
- -- While no one found fault with the Agency'.s identification
of these design obstacles in the adapti-on of trap technology to
heavy-duty use, some of the manufacturers strongly disagreed
with the Agency's conclusions that these obstacles are not
insurmountable and traps would be technically feasible by the
proposed model year (1990). However, none presented spec.ific
, data or engineering analysis to demonstrate a LD/HD difference
to be an insurmountable obstacle. The views of the NPRM were
further reinforced by a document prepared for the Agency, by
Energy and Resource Consultants (ERC), an independent
contractor.[6] This report concluded that light-duty trap
technology can be adapted to heavy-duty use with additional
development time beyond the effective light-duty trap standard
date; line-haul trucks require an extra 3-4 years, as their
operating conditions are the most dissimilar to light-duty
conditions, and the light heavy-duty vehicles whose operating
conditions are more closely related, require only 1-2
additional years at the most. Thus, the extrapolation
contained in the NPRM should be retained. Manufacturers'
heavy-duty test data are examined in the following heavy-duty
trap status section.
c. Heavy-Duty Trap Status
The heavy-duty engine manufacturers' trap development
results examined in the NPRM were obtained from comments the
manufacturers submitted to EPA in 1982 following the initial
1981 particulate NPRM (46 FR 1910) and also from ensuing
meetings between representatives of HDD manufacturers and EPA
staff. The latest comments received in response to the recent
NOx/particulate NPRM added very little test data to what was
evaluated in the NPRM. The following paragraphs will review
and examine the current status of heavy-duty trap development
work as reported by the manufacturers.
-------
2-58
GM submitted a summary of trap development and testing
performed from 1981 through 1983 on HDDEs, much of which had
previously been submitted to the Agency for review. The
successful accumulation of an additional 70.000 kilometers on a
dump truck equipped with a one-piece monolith trap on a
4-stroke turbocharged 8.2L diesel engine was the only new HDDE
testing information received from GM. While GM's statement
that the accumulated mileage (80,500 Km total) is short of the
expected service life of this type of vehicle and the driving
cycle followed was not representative of actual conditions is
correct, trap feasibility in some future year does not require
that traps be fully developed today. In this light, the Agency
views this latest test result as extremely promising. At this
stage in the design of traps, failures are expected; there is
sufficient time to work out trap durability and regeneration
control problems. Despite this, GM feels that traps are
infeasible for production release for the 1990 or 1991 model
year; GM refuses to commit itself to the feasibility of traps
in the forseeable future.
Other manufacturers commented on the feasibility of traps
based on experience in their heavy-duty trap programs. Due to
their laboratory and field testing results, during the last two
years, International Harvester- is quite pessimistic about the
feasibility of traps, with durability being the main design
problem. Field experience, to date, has involved durability
testing with three types of traps on a 6.9L light heavy-duty
engine. Yet despite failures due to inadequate regeneration,
IHC is willing to work towards a trap standard in the 1991/1992
time frame. Mack also is not confident that the durability of
trap systems will be assured. Although Mack expects
regeneration and its control to be feasible, trap durability
remains too much of an unanswered question for Mack to state a
definitive view on trap feasibility. Current work is aimed at
accomplishing regeneration in actual vehicle use; initial
results produced over 6,000 miles of successful operation.
Caterpillar believes that trap technology may not be available
for production by the proposed 1990 model year; however,
Caterpillar did not mention a feasible implementation date
beyond 1990.
As indicated in its comments, Cummins is at an early stage
of heavy-duty trap development. If EPA commits itself to
reassessing the technical feasibility of a trap standard by
December 31, 1987, Cummins would feel comfortable with a 1992
particulate standard of 0.25 g/BHP-hr. However, Cummins added
the caveat that it does not envision traps by 1992. Even
though Volvo White considers current HD trap technology to be
virtually non-existent, it believes that trap 'technology will
be available and qualified by 1991 (as will be discussed below,
this is conditional on the control of sulfur in diesel fuel).
-------
2-59
Daimler-Benz, with the furthest developed heavy-duty trap
program, was the sole HDE manufacturer to agree with the NPRM's
proposed 0.25 g/BHP-hr trap-based particulate standard date of
1990 model year implementation. (This also is conditional on
fuel sulfur control, in addition to an allowable maintenance
condition discussed below.) As described in its 1982 comments
to the Agency, Daimler-Benz is concentrating on the development
of a trap made of wound ceramic fiber. The latest comments
indicate that considerable development progress has been made
in the last two years. Still, much development' remains,
including the optimization of trap design and increasing the
trap durability through an optimized regeneration system.
Current results of urban bus applications of the traps show a
minimum trap service life of 100,000 miles, and a maximum
service life of less than 150,000 miles. With these
encouraging test results at this stage in the design of traps,
the Agency sees no reason that the allowable maintenance
interval of 150,000 miles is not feasible for traps, as
Daimler-Benz indicated in its comments. As EPA has stated
previously in relation to trap feasibility, there is sufficient
time to work out trap durability problems.
The Manufacturers of Emission Controls Association (MECA),
whose member companies are supplying the trap materials being
tested by the HDDE manufacturers, strongly supported the
feasibility of trap-based standards. Although recognizing that
development work remains, MECA stated that a trap-based
standard is achievable, citing worldwide test and development
work by its member companies.
Overall, progress in heavy-duty trap development has not
matched that in the light-duty area over the past two years.
Much of this difference, however, can be attributed to the lack
of a firm target, which can only be a promulgated standard.
While significant steps still need to be accomplished in the
heavy-duty area, the finding that light-duty trap technology
can be extrapolated to heavy-duty engines and thus, traps are
feasible for future heavy-duty usage, remains essentially
unchallenged. The key issue is actually leadtime, which will
be addressed further below.
One issue not considered in the NPRM, and which should be
addressed here was that raised by Daimler-Benz, Volvo White,
and several other manufacturers concerning diesel fuel sulfur
content and its relationship to the heavy-duty engine
environment. Daimler-Benz was very concerned about
trap-plugging by non-regeneratable particulate matter; using a
European diesel fuel, Daimler-Benz found an average of 35
percent of the non-carbon deposits left after regeneration to
be sulfates. While none of the other commenters discussed trie
-------
2-60
issue of trap-plugging by sulfates, Caterpillar and IHC
expressed concern that high sulfate emissions resulting from
high sulfur fuel will make up a significant portion of EPA's
proposed 0.25 g/BHP-hr standard and possibly exceed it. A
reduction in the sulfur content of diesel fuel was recommended
by Daimler-Benz, IHC, Mack, and Volvo White; failing this, or
as an interim step, the manufacturers recommended that EPA
should adopt a correction factor as part of its particulate
test to account for the sulfur portion of the particulate
emissions.
The data available are not sufficient to allow a full
analysis of this issue at this time. Not enough is known about
the Daimler-Benz trap to understand why sulfate plugging is a
_problem there and not elsewhere and what, if any, solutions are
possible short of reducing the sulfur content of diesel fuel.
The comments of other manufacturers presumably apply to
catalyst substrate traps, which generally showed the same
problem on light-duty diesels. (Mercedes* trap is the
exception to this.) As this is not the only trap design, or
even that believed to be the most feasible (which is generally
thought to be burner or fuel additive regenerated), its
elimination from consideration may not affect overall
feasibility. Also, while limited areas in California require
low sulfur diesel fuel, the cost of such control on a
nationwide basis has not been determined and would require
significant study.
Given the uncertainty in the relationship between this
issue and feasibility, it should not preclude implementation of
any trap-based standard. However, the Agency is open to
further discussion in this area and will, on its own, be
analyzing the cost of controlling the sulfur content of diesel
fuel in the future.
Many of the comments on the sulfur issue addressed the
measurement of water, absorbed on the sulfate, as particulate
emissions. Their concerns center on the fact that it is very
difficult to reduce the current conversion of gaseous sulfur
dioxide to sulfate (which is only 2-4 percent). As the
particulate standard becomes more stringent, this sulfate,with
its water,comprises more and more of the allowable emissions.
EPA is currently examining a number of different approaches
which can be incorporated into the test procedure to minimize
the measurement of water. Although commenters recommend a
correction factor added to test procedure to counter the
problem, the time constraints on this rulemaking did not allow
sufficient time to determine the optimum approach. Thus, no
such revisions in the test procedures will be made here;
potential changes will be addressed in a later workshop and
-------
2-61
further study of this issue.. With respect to the sulfate
itself, it should be measured as part of the particulate
emitted as it is definitely inhalable and affects human
health. As the typical conversion of sulfur dioxide to sulfate
has been occurring in all previous heavy-duty particulate
measurements, and thus, estimates of trap efficiency,
feasibility is not affected. Feasibility is only an issue when
sulfate is significantly increased by a catalyst, which was
discussed above.
Some concern was also expressed regarding the fuel economy
effects of trap-oxidizer use. A trap fuel economy penalty
incorporates the fuel economy losses that result from an
increase in backpressure and also the increased fuel
consumption attributed to the energy requirements of positive
regeneration. The NPRM analysis cited a two percent fuel
consumption penalty as the worst penalty which would be
observed with HD traps.
Ford argued that the implementation of traps to HDDEs will
cause fuel consumption to increase by about three percent,
approximately two percent of which is due to the backpressure
portion of the penalty (the average of the "clean trap" penalty
of about one percent and the "loaded trap" penalty of about
three percent). A second manufacturer, Cummins, calculated an
approximate fuel economy penalty of 2.6 percent for a 60-liter
trap; this value was not based on actual testing. As repor'ted
in the NPRM analysis, 1.6 percent of the Cummins estimated
penalty is due to the increased backpressure and 1.0 percent
due to burner-initiated regeneration occurring at 100-mile
intervals. The remaining HDDE manufacturers did not comment on
the fuel economy effects of traps. The Department of Energy
(DOE) noted a zero to one percent fuel economy penalty as a
total contribution from the backpressure on the regeneration.
The effect of trap use on the HDDE fuel economy is cf
course dependent on the trap system design, including trap
type, trap size, regeneration type and frequency of
regeneration. Thus, it is reasonable to expect a range of fuel
economy penalities for the industry, assuming a variety of trap
system designs will be used. The one to two percent fuel
economy penalty range documented in the NPRM analysis is
bracketed by the fuel economy losses submitted by Ford, Cummins
and DOE, with the manufacturers' values on the high side and
DOE's range on the low side. EPA's own test data tend to
support the NPRM range.
In 1983, EPA tested a Corning ceramic trap on both 3 ci. s
engine and a bus chassis.[9] The trap caused up to a 2 ceroer-
fuel economy penalty on the bus chassis, but caused no pen* I'./
-------
2-62
at all on the engine. Steadytstate testing of the engine at
high loads, where the effect should be largest, also showed no
effect.
Also in 1984, EPA tested a 400 hp HDDE with a number of
trap designs. Over the EPA transient test, a Johnson-Matthey
trap mounted close to the exhaust of the turbocharger showed no
fuel penalty. A Corning ceramic trap mounted between the
exhaust manifold and the turbocharger showed a penalty of 2.4
percent. The Corning ceramic traps similarly mounted in
parallel showed a 9 percent fuel economy penalty.
The before turbocharger location maximizes the exhaust
temperature at the trap, but also maximizes the fuel penalty as
it directly affects turbocharger effectiveness. This is
evidenced by the fact that two traps in parallel cause a
greater fuel penalty than a single trap. Normally, use of two
traps would reduce backpressure and reduce any fuel economy
penalty effect. However, here the traps are also acting as
heat sinks, and are removing useful energy that otherwise may
be used by the turbocharger. Two heat sinks are worse than
one. It is extremely unlikely that such a design would be used
on a HHDE where fuel efficiency is of upmost importance. Thus,
the 2.4 percent penalty, generated by this research program
aimed primarily at identifying conditions of spontaneous
regeneration, can be taken as a definite upper limit of any
fuel penalty.
Also of importance is trap size; an increase in trap size
would reduce backpressure and reduce the fuel economy penalty.
In costing trap systems in Chapter 3, larger trap sizes were
used than those used in the NPRM analysis or those used in the
EPA tests above. This was done recognizing that even a 0.5
percent decrease in fuel economy penalty would overwhelm the
added cost of the larger trap.
Even so, the trap size projected in Chapter 3 for HHDEs is
not as large as the 60-liter trap used by Cummins on its 270 hp
engine, which showed the 1.6 percent backpressure related
penalty. This sizeable penalty from such a large trap is
somewhat of an anomaly. For example, data supplied by GM for a
much smaller trap, even accounting for the fact that the engine
was smaller, showed backpressure levels one-third to one-half
lower than those resulting from the Cummins trap. This
discrepancy may be due to trap location- or operating
conditions. The GM backpressure levels result from actual
vehicle road tests, while Cummins' value was not from actual
condition testing. Thus, the Cummins trap seems to have caused
an unusually high fuel economy effect and a .5-1.0 percent
backpressure fuel penalty range based on GM's data is not
unreasonable.
-------
2-63
with respect to regeneration, it is possible to directly
estimate the fuel penalty associated with use of a burner based
system. The burner us.ed in the EPA bus testing and that
forming the basis for the burner cost estimate made by Jack
Faucett Associates were rated at 100,000 btu per hour.
Using an estimated burn time of 5 minutes per
regeneration, regeneration frequencies of 100 miles (used in
Cummins test) and 175 miles (maximum of 145-175 mile range used
in GM test), and HDDV fuel economies taken from the MOBILES
conversion factor analysis, [10] the burner related fuel penalty
ranges between 0.2 and 0.5 percent. This is much lower than
the one percent penalty estimated by Cummins.
Thus, overall a 1-1.5 percent fuel penalty would appear
reasonable for a burner based ceramic , trap system or
approximately 0.5 percent attributed to the burner and
approximately 1 percent attributed to trap backpressure.
However, a fuel additive based trap system would not have the
fuel penalty associated with the burner. This type of system
now appears to be among the most promising. Thus, a range of
0.5-1 percent fuel penalty will be used.
Commenters also addressed the potential safety problems
associated with trap usage. The American Trucking Association
stated .that the high temperatures required for uncatalyzed
oxidation of accumulated particles and the possible dangerous
emissions from catalyzed traps are obstacles in the design of
.safe trap oxidizers. Although Cummins did not detail its trap
safety concerns, Cummins did state that significant work is
needed in the safety area prior to the implementation of traps.
EFA does not dispute that the use of trap oxidizers poses
potential safety problems. But the Agency believes that
through careful design of the trap system the associated risk
can be reduced to manageable levels. One example of a safety
design is to monitor the trap temperature to control
regeneration. For a burner system, flame sensors can shut down
the fuel flow if necessary. The trap design costed in Chapter
3 includes a number .of such sensors. As for the danger of
toxic emissions from catalyzed traps, we assume ATA is
referring to sulfate emissions, which are a recognized problem
with catalyzed traps. EPA is not aware of any hazardous
emissions from non-catalyzed traps, except possibly for
catalyzing fuel additives, which would only be introduced by
the engine manufacturer if safe. It is true that work is
needed in the safety area as traps are developed. However, the
two production or production-ready LDD trap systems appear to
be quite safe and no HD/LD differences appear to prevent such
safe design of HDD traps.
-------
2-64
d. Emission Levels
This section deals with comments on the trap-based
standard of 0.25 g/BHP-hr, separating the issue of trap
feasibility, as examined above, from engine-out and trap
deterioration and trap efficiency, as examined here. "The
emission levels specific for the more stringent 1991 model year
bus and 1994 model year HDE 0.10 g/BHP-hr standard are also
examined.
EPA's determination of the design target level generated a
great deal of comment from the manufacturers. Commenters were
critical of the values used in the analysis for the
deterioration factor of engine-out particulate emissions and
also the deterioration factor of the trap-oxidizer. The
comments on the deterioration factor of the engine-out
particulate emissions and also the AQL adjustment factor were
addressed in a previous section and will not be repeated here.
Several commenters disagreed with the Agency's position
that there is no significant deterioration of particulate
emissions with the use of a trap. They claimed that traps do
deteriorate and thus, a multiplicative DF of 1.0 is
unrealistic. Reasons for trap deterioration, according to
Ford, include: micro-cracks resulting from thermal stress and
high temperatures, leakage at the trap end seals due to
warpage, an increase in the soluble organic fraction and
regeneration control system deterioration. In the NPRM
analysis, EPA did not explicitly consider the occurrence of
micro-cracks, leaks, or an increase in the .soluble organic
fraction. All are theoretically possible, but there are no
data to support their likelihood; the present durability data
show no deterioration. Therefore, a trap deterioration of zero
'is not unrealistic at this time. However, even if trap
deterioration were a factor of 1.2, it would affect
feasibility; it would only require traps to be applied to an
additional 3-10 percent of the fleet, depending on the standard
level and model year being considered.
The subject of trap efficiency, the most variable factor
effecting the emission level, was also addressed in the
comments. One commenter (Ford) did not believe that light-duty
truck trap efficiencies necessarily apply to HDE trap
efficiencies, although it presented no analysis to support this
opinion. Another commenter (Cummins) brought up the
possibility that trap efficiency 'depends on the driving cycle
and also the type of particulate matter; data indicated that
trapping efficiencies for the soluble fractions are about 20
percent less than for the dry particulates in a ceramic
monolith trap.
-------
2-65
The NPRM analysis cited an efficiency range from 70-90
percent for the ceramic wall-f-low monolith trap and a range
from 50-80 percent for the wire mesh trap's collection
efficiency. Daimler-Benz's test results show the collection
efficiency for its ceramic fiber wound trap increasing with the
filter loading regardless of the initial trapping efficiency.
(An unloaded trap with a 60 collection efficiency, increased to
80 percent efficiency with 15 percent loading, 90 percent
efficiency with 40 percent loading, and 97 percent efficiency
with 70 percent loading.) GM commented that it hasn't seen the
efficiencies that EPA reported out of its traps. However, at
another point in its submittal, GM stated that "in spite of
repeated structural failures, with ceramic monoliths, we have
continued their development because of their high trapping
efficiency and overall potential once the control problems for
a consistent regeneration are resolved."[11] ..
. Based upon the above, an 80 percent efficiency was chosen
in the NPRM to represent an obtainable .feasible trap efficiency
level in the 1990 timeframe. One commenter disagreed with this
efficiency level referring to testing of a trap-equipped bus
engine conducted by Southwest Research Institute (SwRI) for
EPA.[9] The transient test particulate emissions of the DDAD
6V-71 engine were reduced 61 percent using a ceramic trap over
-the FTP; over a bus cycle, total particulate was reduced 68
percent. This testing was done on an old engine notorious for
a high soluble organic fraction of its particulate, which
explains the low collection efficiencies. Current technology
engines have much lower HC emissions and lower soluble organic
•vl fractions (SOF) which should result in a much higher trap
efficiency as indicated in Cummins' comments that trap
efficiency increases as the SOF decreases. Other testing
conducced by SwRI[12] did result in a higher trap efficiency; a
Cummins NTC-400 engine equipped with a Corning trap was
effective at reducing particulate emissions by 85 percent.
Thus, an 80 percent efficient trap is still reasonable with
respect to a 0.25 g/BHP-hr standard, if not on the low side of
what traps' actual collection efficiency will be.
Applying the trap deterioration factor, SEA adjustment
- factor and the trap efficiency to the engine-out target level,
(0.42-0.54 g/BHP-hr from above), yields an emission level of
0.10-0.13 g/BHP-hr. Thus, at the 0.25 g/BHP-hr standard, traps
will not be required on all engines; the technically -most
difficult applications will be able to be excluded from trap
usage, which is desirable given the new nature of this
technology. With averaging, approximately 70 percent of the
HDDEs will be trap-equipped in order to meet the 0.25
g/BHP-hr. The percentage of the fleet requiring traps should
decrease to approximately 60 percent by 1994, as the engine-out
target level decreases to 0.42 (discussed in Section above);
this assumes a trap efficiency of 85 percent.
-------
2-66
Limiting traps to only highly efficient, ceramic wall-flow
monoliths, even lower levels can be achieved. Thus, assuming
85-90 percent efficient traps, the engine-out target level of
0.42-0.54 g/BHP-hr results in a emission level of 0.05-0.08
g/BHP-hr.which would comply with the 0.10 g/BHP-hr standard.
By 1994, the engine-out target level is projected to be
0.42 g/BHP-hr. Assuming unchanged deterioration and SEA
adjustment factors and 90 percent efficient traps, this results
in an emission level of 0.05-0.06 g/BHP-hr. Under a 0.10
g/BHP-hr standard and with averaging (excluding urban buses),
roughly 90 percent of the HDDEs will be trap equipped.
In commenting on stringent particulate emission standards,
in addition to trap technology, many commenters addressed the
use of methanol fuel in diesel engines as a method for further
reductions of HDDE particulate emissions. Views on this
subject were widely held. NRDC and other environmental groups
believed that EPA should establish both NOx and particulate
standards based upon the use of methanol as a fuel in new
engines and also set regulations to assure the existence of a
supply and distribution for methanol to fuel heavy-duty
engines.
Comments from HDDE manufacturers expressed caution over
the use of methanol fuels. Saab-Scania stated that it was not
prepared to provide methanol-fueled engines in transit buses in
1990 due primarily to the uncertainty of the unregulated
pollutants and their health effects. This comment was fairly
typical of those by other manufacturers addressing this issue;
'concern over the technological aspects of methanol-fueled HDDEs
was a minor issue compared to the potential health risks of
methanol.
New Jersey Transit and other public transit authorities
believed that EPA should analyze and further evaluate the
feasibility of methanol as an alternative fuel, expressing
concern about the difficulties related to storage,
distribution, operating range limitations for vehicles and the
risks of formaldehyde emissions.
While EPA continues to believe in the potential of
methanol in this area, it considers it premature to actually
set standards requiring the use of methanol. Many basic
questions remain to be dealt with before widespread adoption of
methanol fuel will be possible. Therefore, while continuing to
encourage the development of methanol-based technology, EPA is
taking no action at this time on methanol-based standards.
-------
2-67
e. Leadtime
I_B^^^_WM_«^_^««« • —
i. 0.25 q/BHP-hr Standard
In its NPRM analysis of leadtime, EPA concluded that there
appeared to be sufficient time for the manufacturers to design,
develop, and prepare trap oxidizers for 1990 model year HDDEs.
In their comments, all the manufacturers, some to a greater
degree than others, were cautious in predicting a date for
traps to be in production on HDDEs. Only one ' commenter
(Daimler-Benz) agreed with the Agency's proposed implementation
date, albeit conditionally, as discussed above. The other
manufacturers disputed EPA's analysis that traps would be
feasible for 1990 model year application. In their submittals,
the majority of the manufacturers did provide alternative dates
to the proposed 1990 model year effective date. As reviewed,
International Harvester and Cummins indicated a willingness to
.work towards a trap-based standard in the 1991 to 1992
timeframe. Volvo White expressed its belief that traps will be
available and qualified by 1991. Ford did not believe traps
could be implemented prior to the 1991 model year, if then.
The remaining manufacturers of HDDEs were not certain at what
date in the future traps would be available.
In re-examining the necessary leadtime for the
implementation of a trap-based standard, it appears the
effective date should be delayed from the proposed 1990 model
year to the 1991 model year for several reasons. First, there
has been little apparent progress in HD trap technology
development by the heavy-duty industry over the past two years
(Daimler-Benz being the most notable exception). Light-duty
trap technology has continued to progress and insofar as
heavy-duty trap technology is an outgrowth of light-duty
technology, heavy-duty technology has progressed even without
any overt effort by HD manufacturers. However, not all of this
lack of progress over the past two years is recoverable and an
extra year of leadtime would appear reasonable.
Second, "the promulgation of these standards, is somewhat
later than originally anticipated. (March of 1985 vs. late
1984). While not constituting an entire year, the leadtime was
tight to begin with and an extra year is reasonable for this
reason as well.
Third, the Clean Air Act requires that revised heavy-duty
HC, CO and NOx be at least three years apart. While not
applying directly to this particulate standard, it appears
reasonable to follow this approach in this case. The 5
g/BHP-hr NOx standard is being implemented in 1991 and it is
reasonable to have the particulate standard change at the same
time.
-------
2-68
ii. 0.10 q/BHP-hr Standard
In order to comply with the more stringent 0.10 g/BHP-hr"
standard, traps must be 85-90 percent efficient depending on
their engine-out particulate levels. Presently this efficiency
cannot be obtained by all trap designs, and the design of
high-efficiency traps is generally considered to be technically
more difficult than lower efficiency designs.
In their comments on a 0.10 g/BHP-hr standard, most of the
HODE manufacturers argued strongly that this standard was not
achievable. Only Daimler-Benz and Volvo White believed that
the necessary trap efficiencies were feasible, and this was in
relation to the proposed bus standard, as discussed below. The
Engine Manufacturers Association (EMA) and Ford believed that
trap efficiency must be 85 percent to meet a 0.10 g/BHP hr
standard and believed this level not possible. GM added that
EPA disregarded the variability of trap efficiencies in
assuming a 90 percent efficient trap was possible. Aside from
the general comments on trap efficiency/ technical comments did
not address specific difficulties involved in meeting a 0.10
g/BHP-hr as compared to a 0.25 g/BHP-hr standard (i.e.,
obstacles in the way of obtaining a higher trap efficiency).
The majority of the coramenters .felt this level of
particulate emissions was unobtainable for two reasons. The
first being that the required trapping efficiency would not be
possible by the 1991 model year. The other reason was
discussed previously: high sulfate emissions, resulting from
.high sulfur fuel, will exceed a 0.10 g/BHP-hr standard. While
Daimler-Benz shared the concern over the sulfur issue, the HDDE
'manufacturer stated that the 0.10 -g/BHP-hr standard was
obtainable for 1990 model year buses, depending on an
improvement to the particulate measurement accuracy at low
levels. The Manufacturers of Emission Controls Association
also believed that a 0.10 g/BHP-hr standard was achievable for
1990 model year buses.
Trap efficiency may be increased by either employing a
different trap type or by making design changes to a lower
efficiency trap. Of the trap designs currently considered
promising, the ceramic monolith trap is the most efficient -
its efficiency can be above 90 percent. A ceramic trap
efficiency is related to the porosity of its honeycomb matrix;
high porosity results in low efficiency and vice versa.
Engineering challenges that result from a decrease in the trap
porosity (increase in trap efficiency) include faster
backpressure rises, which must be compensated by increasing the
trap size or by more frequent regeneration. The forrer
solution may also solve the potential increase in ash or fuel
-------
2-69
additive accumulation. The latter may not. A larger trap can
create additional design problems itself (greater stresses and
heating requirements for regeneration) as were discussed in the
NPRM analysis of light-duty/heavy-duty design differences.
More frequent regeneration is fairly simple for a burner-based
system, but may be much more difficult for catalyzed or
fuel-additive based systems where naturally occurring
temperatures are relied upon to induce regeneration.
Due to the increased difficulty in designing a
higher-efficiency trap capable of complying with a 0.10
g/BHP-hr standard, the technical feasibility of all 1991 model
year HDDEs complying with this standard is not likely. " By
establishing a 1994 0.10 g/BHP-hr standard, the Agency believes
that the additional three years will allow for the development
of higher efficiency traps with more time to optimize
performance and durability while minimizing cost.
""Theoverall ""difficulty of .achieving these high
efficiencies also depends on the number of engines needing to
be so equipped. It is likely that a number of trap systems
employed to meet the 0.25 g/BHP-hr averaging standard will be
85-90 percent efficient. Others will be less efficient. Thus,
for a few engine urban transit buses, for example, a '0.10
g/BHP-hr standard should be quite feasible. In actuality, few
bus engines are currently marketed in the U.S. GM dominates
the market with its 6V-92TA and 8V-92TA bus engines. Cummins
has sold a small number of its VTB-903 engines in buses in the
past, but is not currently doing so. A small number of
foreign-based manufacturers, such as M.A.N. and Daimler-Benz,
have recently begun to market bus engines in the U.S. Thus,
bus engines represent a relatively small subset of HDDEs.
Developing a trap-oxidizec system for transit bus use may also
be considerably easier than for most HDE applications. An
EPA-sponsored report by Energy Resource Consultants, Inc.[6]
mentions the following reasons this may be so:
1. Durability and reliability requirements would not be
nearly as strict as for most other types of heavy-duty vehicles.
2. Buses have a rather predictable operating cycle, and
and one which includes a great deal of acceleration. The
frequent occurrence of moderate high exhaust temperatures as a
result would help to make a self-regenerating system feasible.
3. Transit buses universally receive regular service,
often on a daily basis.
-------
2-70
Thus, at minimum, bus engines should be no more difficult
to trap-equip than other HDDEs and, at best, could be much
easier to trap-equip. Coupled with their small number,
developing high-efficiency traps for bus engines should be
feasible at the same time as traps are generally employed on
HDDEs, or 1991.
Equipping all HDDEs with high-efficiency traps will
require additional time beyond 1991._ Having buses operating
with such traps will certainly provide useful data, but such
data cannot be employed any sooner than three years after the
buses begin service, since time is required to obtain the data
and design and tooling must also be performed. Of more use
will be durability data generated on prototype non-bus HDDVs
equipped with high-efficiency traps after the bus engine
designs have been set, but prior to bus introduction.
Providing only two years between standards can reasonably be
ruled out due to the need to apply such traps to line-haul
"HDDEs, which have very long lives and which require extensive
durability data. The argument can be made that three years
should be sufficient to incorporate such durability data. As
this also coincides with the Act's requirement for HC, CO, and
NOx standards, it appears the most reasonable interval time as
well.
.C. Conclusions
1. Near- and Mid-term NOx and Particulate Standards
As a result of the proceeding analysis of the comments,
EPA has concluded that the proposed standards of 6.0 g/BHP-hr
NOx and 0.60 g/BHP-hr particulate are technologically feasible
and that the appropriate date for implementation of these
standards is the 1988 model year.
EPA has also concluded that engine-out emission standards
of 4.0 g/BHP-hr NOx and 0.40 g/BHP-hr for HDDEs are not
technologically feasible using any known emission control
technology. Information provided in the comments has, however,
lead EPA to modify the NPRM analysis and conclude that an
engine-out NOx emission standard of 5.0 g/BHP-hr is
technologically feasible by the 1991 model year. The lowest
feasible engine-out particulate level, given the 5.0 g/BHP-hr
NOx standard, appears to be 0.50 g/BHP-hr.
With respect to fuel economy, the 6.0/0.60 standards are
expected to cause a 0-2 percent fuel economy penalty in the
near term and that this penalty will be erased by 1991. The
NOx standard of 5.0 g/BHP-hr is expected to cause approximately
a 1 percent penalty initially, decreasing to approximated 1/2
percent in a few years.
-------
2-71
2. Trap-Based Particulate Standards
As a result of the proceeding analysis of the comments and
additional information, the Agency concluded that trap
technology is feasible for heavy-duty diesel engine
application. This conclusion was extrapolated from the status
of light-duty trap technology and the design effort necessary
to adapt this technology to heavy-duty usage. Light-duty traps
have been proven to be a feasible control of particulate
emissions from light-duty vehicles. Although 'conditions
specific to the HDDE environment require considerable
development in order to apply LD trap technology to HD usage,
these obstacles are not insurmountable and with adequate
engineering effort traps should be a feasible control method of
particulate emissions from heavy-duty vehicles.
EPA has also concluded that the 0.25 g/BHP-hr trap-based
particulate standard should be feasible for 1991 model year
'HDDEs; the 0.10 g/BHP-hr trap-based particulate standard should
be feasible for 1991 model year urban buses and for all 1994
model year HDDEs. The analysis determined that the 1991 model
year HDDE standard, with averaging, will require 80 percent
efficient traps on roughly 70 percent of the fleet. This would
decrease to about 60 percent after the initial years. The '1991
model year bus standard will require the use of nearly 90
percent efficient traps on all buses. The 1994 model year HDDE
standard will require the use of 90 percent efficient traps- on
roughly 90 percent of all HDDEs.
-------
2-72
References
1. "Draft Regulatory Impact Analysis and Oxides of
Nitrogen Pollutant Specific Study," U.S. EPA, OAR, QMS.
2. "Assessment of Domestic Automotive Industry
Production Lead Time of 1975/76 Model Year - Final Report."
Aerospace Corporation for U.S. EPA, OMSAPC, ECTD, December 1972.
3. "Trap-Oxidizer Feasibility Study," U.S. EPA, OANR,
OMS, ECTD, SDSB, Public Docket No. A-83-32, February 1982, .
4. "An Updated Assessment of the- Feasibility of
Trap-Oxidizers," Regulatory Support Document, J. Alson and R.
Wilcox, U.S. EPA, OANR, OMS, ECTD, SDSB, Public Docket No.
A-82-32, June 1983,
5. "Trap-Oxidizer Technology for Light-Duty Diesel
Vehicles: Feasibility, Costs and Present Status," Energy and
Resource Consultants, Inc., Final Report for U.S. EPA.
Contract NO. 68-01-6543, Public Docket No. A-82-32.
6. "Particulate Control Technology and Particulate
Emission Standards for Heavy-Duty Diesel Engines," Energy and
Resource Consultants, Inc., Report to U.S. EPA, Office of
Policy and Analysis, EPA Contract #68-01-6543, December 11,
1984.
7. "Mercedes to Use Traps on 1985 Turbodiesel," Wards
Engine Update, Volume 10, No. 13, July 1984.
8. "'86 VWA Diesels Will Have Traps," Wards Engine
Update, Volume 10, No. 20, October 15, 1984.
9. "Preliminary Particulate Trap Tests on a 2-stroke
Diesel Bus Engine," Ullman, T. L., Hare, C. T. , Southwest
Research Institute, Baines, T. M. , EPA, SAE Technical Paper
Series, 840079, February 1984.
10. "Heavy-Duty Vehicle Emission Conversion Factors
1962-1997" Smith, M.C. IV, U.S. EPA, OAR, OMS, ECTD, .SDSB,
EPA-AA-SDSB-4-1, August 1984.
11. "General Motors' Final Comments on the October 15,
1984 Notice of Proposed Rulemaking with Respect to Gaseous
Emissions Regulations for 1987 and Later Model Year Light-Duty
Vehicles, Light-Duty Trucks, and Heavy-Duty Engines and
Particulate Emission Regulations for 1987 and Later Model Year
Light-Duty Diesel Trucks and Heavy-Duty Diesel Engines,"
Submitted to Environmental Protection Agency, Washington, D.C.,
December 17, 1984.
12. "Heavy-Duty Engine Exhaust Particulate Trap
Evaluation," U.S. EPA, OMSAPC, ECTD, EPA 460/3-84-008,
September 1984.
-------
CHAPTER 3
ECONOMIC IMPACT
This chapter analyzes the costs of complying with the new
NOx and diesel particulate standards in light of the comments
received in response to the NPRM. These comments at times
supported and at times disputed the EPA cost estimates; some of
these comments prompted revisions of the costs given in the
NPRM, and are outlined below.
The chapter begins with a synopsis of the methodology used
in the Draft Regulatory Impact Analysis (RIA) to generate .the
cost estimates in the NPRM. Following the synopsis is the
Summary and Analysis of Comments, which is divided into three
cost sections: LOT, HDGE, and HDDE. Within each section is a
summary of the applicable comments, discussion of how -the
comments compare to the information contained in the NPRM, and
any reanalysis as necessary. Each section closes with a
summary of the final cost and cost-rel-ated values used in the
final economic impact analysis. This is followed by a
discussion of socioeconomic impacts.
I. Synopsis of the NPRM Analysis
This chapter as originally presented in the Draft RIA
examined the compliance costs of the proposed NOx and diesel
particulate standards for LDTs and HDEs. It included the
manufacturers' fixed costs of pre-production (research,
development, and testing (RD&T), including certification
testing), and their variable costs of production (emission
control hardware component costs), as well as user costs of
increased purchase price, fuel economy losses, and maintenance
cost changes. The chapter was divided into two sections which
discussed, respectively, the actual manufacturer and user
costs, and the socioeconomic impacts of such costs. The first
section is the more lengthy one, and received the main bulk of
the comments. It is summarized below. The socioeconomic
impact section," which included manufacturer, regional, and
national effects on sales, cash flow, employment, balance of
trade, and consumer prices received comments on two issues
only, and therefore need not be reviewed in full.
Commenters on costs focused on alternative values to the
costs derived by EPA rather than on the methodology used,- and
therefore the methodology is described here only briefly. Any
interested parties may consult the Draft RIA for more complete
information on the cost derivation methodology and actual cost
values which were presented in the NPRM.
-------
3-2
A. Cost to Manufacturers
In EPA's analysis, manufacturer costs for each of the
vehicle/engine groups — LOT, HDGE, and HDDE — included the
fixed 'costs of RD&T and the variable costs of hardware. Fixed
costs were determined by estimating the number of
recalibrations, design modifications, and the amount of new
testing necessary to convert present systems to those which
could meet the standard. Numbers of calibrations needed per
engine family were combined with numbers of engine families
needing the work, estimated hours of effort per calibration,
hourly rates for labor, overhead and parts, and a 10 percent
'contingency factor to derive a dollar value for total
recalibrations. Similar estimates were made of the time
-necessary for redesign and for completely new, general system
designs, such as that required for particulate traps.
Testing costs to prove mechanical integrity were based on
miles of testing necessary, average speed, and hourly rates for
labor and overhead; such test costs were shared with other
testing programs when applicable. Certification testing costs
included the same type of mileage accumulation costs, as well
as fixed costs of $1,500 per emission test for LDTs and $2000
per emission test for HDEs.
For LOT, HDGE, and 1987 HDDE proposed standards, it was
assumed that these fixed costs would be incurred in the fcwo
years prior to implementation of the standards; for the 1990
HDDE standard, four years were allotted due to the longer
development time needed for trap-oxidizer systems. The sum of
all these costs was apportioned over five model years for LDTs,
and three model years for HDEs (due to introduction of the
second set of HDE standards after three years) . Costs were
presented in both undiscounted and discounted forms.
Discounted costs were calculated at a 10 percent discount rate
to the first year for which the standard was applicable (1987
or 1990). These costs were then spread over projected sales to
determine an average cost per vehicle or engine due to RD&T.
Variable costs to manufacturers arose from the addition of
new hardware and, in some cases, credit was taken for the
removal of old hardware components. For LDTs these component
costs were developed using the Rath and Strong methodology[1]
as discussed in the Draft RIA and include overhead and
manufacturer profit. For HDGE and HDDE NOx and non-trap
particulate control on HDDEs, component costs were developed
from costs for similar pieces of equipment on current engines,
with inclusion of factors for different material costs due to
different component sizes. Particulate trap costs were taken
directly from the Diesel Particulate Study.[2] These component
cost estimates were combined with projections of the technology
-------
3-3
changes which would be necessary to meet the new standards,
rr.arket snares of various technology mixes and vehicle/engine
types, and projected sales, to develop both per vehicle/engine
and aggregate manufacturer costs for hardware. Hardware costs
were combined with RD&T costs, with appropriate discounting at
10 percent, to determine total manufacturer costs.
B. Cost to Users
Costs to users were based on increases in first cost at
the retail price equivalent (RPE) level and additional
operating costs due to changes in fuel economy and
maintenance. The first price increase per vehicle includes 'the
average hardware cost for that vehicle's new technology
application and the per vehicle share of RD&T, which is
apportioned over the three or five years after implementation
of the standard as discussed above.
The lifetime cost changes per one percent of fuel economy
change were calculated from fuel price, average base fuel
economy and lifetime mileage per vehicle or engine category,
using a 10 percent discount rate. Overall fuel economy changes
expected were estimated in the Technological Feasibility
Chapter according to the types-of technology necessary to meet
the standard. Lifetime fuel economy costs due to the standard
could thus be calculated by multiplying cost per one percent
change by the amount of change expected.
Maintenance costs were determined from any additional
maintenance operations deemed necessary for the new technology,
the expected number of additional maintenance operations per
lifetime, and the cost per occurrence. These costs were then
discounted at the usual 10 percent rsts, from point of
maintenance to point of sale. The costs were then apportioned
two different ways. In the first case, they were apportioned
.over just the engines requiring the new technology and the
corresponding maintenance, and in the second case the costs
were apportioned over all engines. In appropriate cases
credits were taken for maintenance which would be reduced,
using the same methodology. Maintenance costs, either negative
or positive, were added to fuel economy costs and first price
increases to give the total lifetime user cost of the standard.
Aggregate costs to the nation, including those to both
manufacturer and user, were then calculated for each vehicle or
engine group. Hardware costs plus operating costs of fuel and
maintenance were multiplied by number of vehicles affected,
according to future sales projections. As noted above, a
5-year period was used for LDTs and 3-year periods for HDEs.
These values were discounted to the proposed year of
implementation, and added to discounted RD&T costs, to yield
-------
3-4
the net present value aggregate cost in the year the standards
begin. Costs to the nation were -also expressed in terms of per
vehicle investment in 1987. All values used throughout the
chapter were 1984 dollars.
II. Summary and Analysis o£ Comments
A. LPT NOx Standard
1. 1988 NOx Standards
a. Cost to LPT Manufacturers
Comments on the manufacturer costs attributable to the new
LDT NOx standards were neither numerous nor lengthy, with only
two manufacturers estimating retail price increases due .to
hardware and RD&T. The price increases in the comments were
stated without derivations, and the methods used by EPA to
estimate the costs given in the NPRM were not challenged. The
manufacturer and EPA final costs, however, were not based on
comparable fleets. Manufacturer costs were based on only those
vehicles requiring new technology, while the EPA final costs
developed in the NPRM were expressed as an average over every
LDGT or LDDT, including those vehicles which will already have
the hardware in place prior to implementation of the standard.
EPA developed an average per vehicle cost for a LDGT requiring
new technology in the Draft RIA, as part of the process of
developing the fleet average LDT costs. This cost ($140) was
presented in the Draft RIA. The EPA and manufacturer cost
'estimates are as follows: ....
Retail Price Increase Per LDT
Draft Analysis $35 LDDT Average
, , ' $44-87 LDGT Average
$140 LDGT with new technology
Chrysler . $80 _ _ LDT with new technology
Toyota " $100-250 LDTt with new technology
As can be seen, when the same figures — cost per vehicle
with new technology — are compared, those given by the
manufacturers are comparable to those estimated by EPA in the
NPRM. Therefore, EPA sees no need, to alter its analysis of the
LDT component cost values or RD&T costs which were used to
develop the average per vehicle values presented in the NPRM.
i. Fixed Cost
In developing the costs for the final rule, the original
allocation for RD&T costs estimated by EPA for this standard
will 'be used here, but shifted one year to concur with the
-------
3-5
shift in the year of introduction of the standard. This amount
is the sum of the RD&T costs for LDGTs and LDDTs, and totals
$26,970,000.
ii. Variable Cost
The average costs have been updated to reflect
manufacturer comments on projected technology mixes received in
response to the NPRM. Present technology mixes are taken from
1985 model year certification data, which provides manufacturer
sales projections by engine family and, hence, by emission
control technology type. The manufacturer data from comments
and the 1985 certification data provide the basis for revision
of the projected technology mixes and subsequent revision of
the average LDGT hardware costs, which are calculated in the
same manner as used in the Draft RIA. The projected mixes were
in most cases confidential on a manufacturer-specific basis,
and are included here only in general form.
The new present (1985) and projected technology
application mixes according to LDGT engine size are given in
Table 3-1. The same information for all LDGTs combined is
presented in Table 3-2. The 1985 model year mixes were
converted to projected 1987 model year (pre-standard) mixes by
applying technology changes on specific engine families as
indicated by manufacturers in confidential comments to the
NPRM. As noted in the NPRM, the comments verify that even with
no increases in the stringency of the NOx standard, there is a
clear trend away from oxidation catalyst systems to
three-way-catalyst systems on LDTs, apparently for reasons such
as improved driveability and fuel economy. The trend for LDGTs
is predominantly toward three-way closed-loop systems as
opposed to three-way open—loop systems. Therefore, this
analysis projected that for the 1988 model year (the first year
of the new standard), all remaining oxidation catalyst systems
and all three-way open-loop systems will convert to three-way
closed-loop technology.
- The overall hardware cost for each vehicle undergoing a
technology change, according to number of cylinders in the
vehicle, is given in Table 3-3, which is a summary of Tables
3-4 through 3-6 in the Draft RIA. Costs are derived by
subtracting the costs of hardware components removed from the
costs of hardware components added; estimates of these costs
were originally determined using the Rath and Strong
methodology,[1] and are unchanged from the Draft RIA.
The hardware costs as presented in the Draft RIA and the
projected change in the technology mix, as discussed above, are
combined to give an average hardware cost upon implementation
of the 1988 model year standard for LDGTiS and for LDGT2s
as follows:
-------
3-6 -
Table 3-L
Light-Duty Gasoline Trucks
Percent Technology Usage 3y Model Year and Engine Size
Oxidation
Three-Way
Open-Loop
Three-way
Closed-loop
Three-Way
Open-Loop
plus
Oxidation
LDGTL:
1985 MY*
4-cylinder
"6-cylinder ~ '-"I. ~~
ALL
1987 MY projected**
4-cylinder
6-cylinder
ALL
1988 MY projected**
4-cylinder . • 0
6-cylinder • £
ALL . • 0
0
0
0
0
0
0
0
0
0
38.5
0
3875
0
0-
0
0
0
0
0
0
0
Three-Way
Closed-Loop
plus
Oxidation
1.8
11.77
13.5
LDGT2:
1985 MY*
6-cylinder
8-cylinder
ALL
3.3
34.0
3773
1987 MY projected**
6-cylinder
8-cylinder
ALL
1988 MY projected**
6-cylinder . 0
3-cylinder £
ALL . 0
0
0.8
0.8
0
0
0
0
18.4
T874
3.3
53.2
T675
0
8.7
877
0
8.7
8.7
0
0
0
•**
Based on manufacturers' confidential sales projections'.
Based on manufacturers' confidential comments to the NPR4.
-------
3-7
Table 3-2
Ligr.t-Duty Gasoline Trucks
Percent Technology Usage by Model Years for All LDGTs
Percent Technology Used
Model Year
1982*
1983*
1984*
1985*
1987**
1983 and
later1
No
Catalyst
' 4.2
1.4
0.0
0.0
0.0
• 0.0
Oxidation
Catalyst
2
3
91
80
58.0
43.3
36.9
0.0
Three-way
Open-Loop
0.2
0.0
11.0
4.9
3.6
0.0
Three-way***
Closed-Loop
4.4
18.3
31.0
51.8
59.5
100.0
**
*»*
Based on confidential sales projections provided by
manufacturers as part of the certification process.
Projected based on confidential manufacturer comments to
the NPRM.
Also includes three-way closed-loop plus oxidation
catalyst systems (see Table 3-1).
-------
3-3
Taole 3-3
Lignt Duty Gasoline Truck
Emission Control System Hardware
Cost Per Vehicle with New Technology
Technology Change Engine Size
From To 4CYL 6CYL 8CYL
Oxidation Three-way Closed Loop $106 • $133
Three-way Three-way Closed Loop
-------
3-9
It is projected that 48 percent of all LDGT-s will
require technology changes between 1987 and 1988.
First, 4-cylinder LDGTiS, representing 21.2
percent of all LDGTt sales, are converted from
oxidation catalysts to three-way closed-loop
catalysts. At $106 per conversion, the contribution
of 4-cylinder engines to the LDGT, hardware cost
is: 0.212 X $106 - $22.47
Six cylinder LDGTts, representing 26.8 percent of
all LDGTis, are changed from oxidation catalysts
to three-way closed-loop catalysts, at $133 "per
vehicle. The 6-cylinder contribution to the LDGTt
hardware cost is then: 0.268 X $133 = $35.64
The remainder of the LDGTi market, 52 percent, is
projected to already have the required technology
(38.5 percent three-way closed-loop catalysts and
13.5 percent three-way closed-loop plus oxidation
catalysts) in place for the 1987 model year. (In
fact, most of these vehicles, 49.4 percent of the
market, already have the hardware on the 1985 model
year vehicles.) No. costs are incurred for these
vehicles.
The average hardware cost per LDGTi in the fleet
is the sum of the contributions by the 4- and
6-cylinder engines and is: $22.47 + $35.64 = $58.11
or approximately $58.
0 This cost is $121 per vehicle when applied only to
uuGse ULJ\J ± \ 3 requiring new u€CnHGj.ogy in mo w£ j.
year 1988.
The same methodology can be used for LDGT2s:
°, For the 3.3 percent of LDGT2s which are 6-cylinder
engines and which are converted from oxidation
catalyst to three-way closed-loop catalyst
technology, the cost is $133 per vehicle, which
results in a contribution of: 0.033 X $133 = $4.43
0 For the 15.6 percent of the LDGT2 market which, are
8-cylinder engines with oxidation catalysts and
which will be changed to three-way closed-loop
catalyst systems at a cost of $157, the resulting
8-cylinder contribution is: 0.156 X $157 = $24.49
-------
3-10
0 The 8-cylinder engines also have some open-loop
three-way systerr.s, which will go to closed-loop in
model year 1988. Of"the LDGT2 market, 9.5 percent
will need new closed-loop control (8.7 percent from
three-way plus oxidation catalysts to three-way
closed-loop plus oxidation catalysts, and 0.8
percent from three-way to three-way closed-loop
catalysts) at $65 per vehicle, which makes this
portion of the 8-cylinder contribution: 0.095 X $65
« $6.18
0 The remainder of the LDGT2 market, 71.6 percent,
- - -already will have three-way closed-loop catalyst or
three-way closed-loop plus oxidation catalyst
systems in place by model year 1987, and thus will
incur no costs.
0 The average hardware cost per LDGTZ in the fleet
^ ^ is the sum of the 6- and 8-cylinder contributions
-~~"~and is: $4.43 * $24.49 + $6.18 - $35.10, or about
$35.
0 This cost, when distributed only over those vehicles
requiring new technology, is about $123 per LDGT2.
The average per vehicle hardware costs of $58 for LDGT(s
and $35 for LDGT2s derived above include the total costs for
hardware components added (three-way catalysts, feedback
carburetor modifications, and/or closed-loop control), with a
credit for those components removed (oxidation catalysts and/or
open-loop control), averaged over the entire fleet of vehicles
in that LDGT category. However, the complete cost of such
hardware should not be applied solely to the more stringent
NOx standard, since the manufacturer also derives other
significant benefit from its application. This is indicated by
the fact that manufacturers have already converted much of
their fleet from oxidation to three-way catalyst systems, and
have stated in their comments that they plan to continue this
trend, even though it is not necessary from an emissions
control standpoint under the current 2.3 g/mi NOx standard.
The application of this more costly technology prior to the
implementation of the stricter standard clearly indicates
benefits to the manufacturer, which include improved fuel
economy and driveability as well as parts consistency with
their light-duty vehicles. This .parts consistency leads to
greater economic efficiency and lower total costs. It is
difficult to quantify the precise value of these benefits to
the manufacturer; however, it is clear that they are
significant. Absent any more precise information, EPA has
applied 50 percent of the costs to the implementation of the
standard, and 50 percent to the other benefits which will be
-------
3-17
Table 3-8
Total LPT Manufacturer Cost
1986
1987
1988
1989
1990
L991
1992
TOTAL
Undiscounted
RD&T Hardware
$10,900K
$16,070K
-- . $92,350K
$94,200K
$93,67<3K
S97.110K
$98,960K
$26,970K $476,290K
$503,260K
Discounted"
RD&T Hardware
S13.190K
$17,580K
, — $92,350K
$85,640K
$77,410K
$72,960K
$67,590K
$30,870K $395,950K
$426,820K
Discounted at 10 percent to 1988.
-------
3-18
Table 3-9
Light-Duty Trucks F'irst Price Increases
Vehicles Requiring New Technology
LDGT :
4-Cylinder
6-Cylinder
8-Cylinder
8-Cylinder
Three-way Closed Loop from
Oxidation Catalyst (LDGT,)
Three-way Closed-Loop From
Oxidation Catalyst (LDGTi » ,)
Three-Way Closed-Loop From
Oxidation Catalyst (LDGT2)
Three-Way Closed-Loop From
$109
$136
$160
$ 68
LDDT:
LOOT,
LDDTz
Three Way Open Loop (LDGTZ)
First Time Application of EGR
Electronically Controlled EGR from EGR
Average for All Vehicles
LDGT»
LDGT 2
LDDT,
LDDT2
All LDTs
$ 22
$ 44
$ 31
$ 20
$ 22
$ 44
$ 28
-------
3-19
rray experience a gain of up to 8 percent. This was seen in
comparisons of 1985 certification data of matched pairs of
Federal and California vehicles. For fuel economy changes
which may occur, the costs remain as in the Draft RIA at $51
for LDGTs and $41 for LDDTs lifetime cost per affected vehicle
per one percent change in fuel economy, either greater or
less. When apportioned over all vehicles, the cost is $21 per
LDGT and $41 per LDDT, or $24 per LOT per one percent change in
fuel economy. It is expected that on a fleetwide basis there
may be a slight gain in fuel economy; however, for costing no
fuel gain was included.
iii. Total Cost to Users
To summarize, purchasers of LDTs can expect to pay an
average of $28 more for 1988 model year LDTs for the emission
control improvements as compared to 1987 LDTs. In the case of
fuel economy increases or decreases, LDGT users can expect a
$51 change in lifetime operating cost per one percent change in
fuel economy, while LDDT users can expect a $41 change. A
slight gain in fuel economy is expected, but is not included in
total cost.
2. Aggregate Costs for the 1988 LPT NOx Standard
The aggregate cost to the nation of complying with the
1988 Federal LOT NOx emission regulations consists of the sum
of fixed costs for RD&T and new emission control hardware. No
changes in maintenance or fuel economy costs are expected.
These costs are calculated based on sales projections for the
5-year period following introduction of the standard. These
sales projections were shown in Table 3-4.
The various costs associated with this rulemaking action
will occur in different periods. In order to make all costs
comparable, the present value at the start of 1988 has been
calculated based on a discount rate of 10 percent. The
.calculations were shown earlier in Table 3-8. The aggregate
cost of complying with the new regulations for the 5-year
period is estimated to be equivalent to a lump sum investment
of about $427 million (1984 dollars) made at the start of
1988. When amortized over the projected sales for the
five-year period, the value is $28 per vehicle at the time of
purchase.
-------
3-20
B. HDGE NOx Standards
Comments on the costs of the proposed HDGE NOx standards
were received from two of the three major manufacturers of
KDGEs and were not highly detailed. Chrysler gave a cost
estimate only for the later standard, with no estimate for a
fuel economy penalty. General Motors stated that, "the
predominant HDGE costs associated with the more stringent NOx
standards proposed by EPA would be an increase in fuel
consumption." Ford comments did not discuss the costs of the
standard. Therefore, any cost revisions below are based on a
reanalysis of the control technology necessary, which the
manufacturers' comments did discuss, rather .than on concerns
for cost estimates for specific components of the control
technology.
Before beginning this reanalysis of the costs to comply
with the 1988 and 1991 HDGE NOx standards, a brief discussion
of - HDGE certification data and options and potential
certification approaches is necessary. First, as of February
1985, the three major HDGE manufacturers had certified a total
of seven families. This is a decrease of eight families from
1984, brought about by IH leaving the market completely,
Chrysler dropping one family, and GM and Ford each combining
two families which were previously separate. However, due to
the split class HDGE emission standards beginning in 1987, the
number of HDGE families is projected to increase from 7 to J.O
or 11 even though no new engine offerings are expected. For
simplicity, and since all HDGEs will have to meet the same NOx
standards, this analysis will assume that HDGE sales are spread
evenly among the 11 families. This allows fixed costs to be
assigned on a per family basis and spread over the entire
fleet, without having to assign specific fixed costs to
specific families for the sake of production-weighting the
fixed cost impacts. As will be seen later, in the long term
this introduces no error into the per vehicle cost.
Second, it is worth noting that beginning in 1987,
manufacturers may exercise the option to certify their HDGVs of
up to 10,000 Ib GVW as LDTs. While this option also exists for
the 1988 model year, this analysis does not evaluate that
possibility. Presumably it would be more expensive on a per
engine basis to meet the LOT NOx standard (probably requiring a
three-way catalyst with closed-loop control) than it would be
to meet the 1988 or 1991 HDGE standards. -Therefore, if
manufacturers choose to exercise this option in 1988, it would
be based on their belief that other perceived or intangible
benefits are worth any extra costs.
Third, any analysis of costs to meet the 1988 or 1991 HDGE
NOx standards must be placed in the proper context by reviewing
current HDGE NOx certification levels. The certification data
-------
3-21
presented in the HDGE technological feasibility analysis
indicates chat none of the families certified in 1985 .T.eet the
1988 6.0 g/BHP-hr NOx standard, even though two configurations
within these families do meet the 1988 standards anc one of
these cwo configurations meets the 1991 standard. For purposes
of cost estimation, this analysis will assume that no current
HDGE families meet the 1988 NOx standard of 6.0 g/BHP-hr.
However, several will be very close.
1. 1988 NOx
a. Cost to HDGE Manufacturers
• i. Fixed Costs
As noted above, no comments specifically addressed the
issue of manufacturers' cost of the intermediate HDGE NOx
standard. •-However, a reanalysis of cost has been done to
reflect changes in the EPA projection of the control technology
necessary to meet the standard, as discussed in Chapter 2,
Technological Feasibility. These new cost estimates are
outlined below.
The costs originally estimated for RD&T included
recalibration of the fuel, ignition, and EGR systems as well as
certification costs. In total, these recalibrations amounted
to $39,600 per engine family based on three calibration
combinations at six person-weeks of effort each. Certification
costs in the Draft RIA were $192,170 per engine family, based
on one durability and three data engines per family.
Ths final cost estimate includes these costs, plus costs
for evaluation and recalibration for improved secondary air
management, redesign of the combustion chamber, and
emission-related improvements of the intake manifold.
Secondary air management recalibration is expected to require
about the same level of effort as the fuel, EGR, or ignition
system, or $13,200 per engine family. Redesign and testing of
the combustion chamber is expected to consist of a redesign of
the cylinder head and was estimated in the Draft RIA for the
proposed 4.0 standard to cost $306,900 per engine family,
(including a 10 percent contingency factor) . This value is
used here. Finally, enhancement of the intake manifold is
estimated to require about five times the level of effort of
any. of the above recalibrations, an amount of $66,000 per
engine family.
As was discussed above, it is now projected that the
major manufacturers will certify a total of 11 HDGE families in
1988. The three tasks originally presented in the Draft RIA
are expected to be necessary for all 11 engine families. Total
cost is thus $580,800 for recalibration of EGR, fuel, and
ignition systems.
-------
3-22
Of the 11 engine families receiving this recalibration, 8
are expected to be able to rr.eec the 6.0 standard without
additional changes. This is based on a review of current NOx
certification levels which indicates only three families have
NOx emission levels above 8.0 g/BHP-hr. The other three HDGE
families may require some or all of the additional work listed
above: secondary air recalibration, combustion chamber
redesign, and intake manifold improvement. As given in the
Draft RIA, RD&T costs for applying these improvements to each
of the remaining engine families are $13,200 for the secondary
air, $306,900 for the combustion chamber, and $66,000 for the
intake manifold for a total of $386,100 per family. Assuming
that all three families do all the work,, this totals an
additional $1,158,300.
Certification must be conducted for all 11 engine
families. Using the $192,170 certification cost per family
presented in the Draft RIA, certification costs total
.$2,113,870. The total fixed costs to meet the 1988 HDGE NOx
standard is the sum of the development and certification costs
or about $3,708,000. The separate components of these costs
are detailed in Table 3-10.
ii. Variable Cost
EGR is the major NOx emission control component expected
on HDGEs. A review of the 1985 certification records indicat.es
that six of the seven HDGE families currently have EGR
installed. One family, representing about 4 percent of sales,
would have to install EGR to meet the 1988 HDGE NOx standard.
In a 1980 study completed for EPA, HDGE EGR was estimated to
cost $9.36 at the vendor level (1977 dollars).[10] When
adjusted for inflation to 1985 dollars using the new car CPI
(1.43) and accounting for manufacturer and dealer overhead and
profit (1.29), [9] HDGE EGR is estimated to have a retail price
equivalent of $17.27. When spread over all the vehicles in the
fleet, this averages $0.69 per vehicle.
In addition the recalibration discussion above indicated
that 3 families may need additional work for combustion chamber
modifications and intake manifold 'improvements. Redesigned
hardware may be necessary for the three HDGE families needing
this work. In the long term, the redesigned parts could
presumably cost the same as those being replaced, but a
conservative approach is taken and $10 is assigned per
redesigned engine. When this cost for work on three engine
families is spread over all HDGEs, the hardware cost per
redesigned engine due to the &.0 standard is $2.73. When the
EGR cost is added to the redesigned component cost, the total
hardware cost sums to $3.42 per HDGE.
-------
3-23
Table 3-10
Summafy of 1988 HDGV NO* RD&T Costs
Families Cost per
Category Affected Family Total
Fuel, ingition and EGR 11 $39,600 $435,600
recalibration
Secondary air management, 3 $386,100 $1,153,300
combustion chamber redesign,
and intake manifold mods.
Certification 11 $192,170 $2,113,870
$3,707,770
-------
3-24
iii . Total Manufacturer Cost
The total manufacturer cost of the 6.0 HDGE NOx standard
is the sum of the RD&T cost and the hardware cost for the
engines produced in the three model years immediately following
introduction of the standard, all discounted at 10 percent to
1988. Projected sales have been updated to reflect information
in Reference 3, and are presented in Table 3-11. These sales
figures have been used to generate the total manufacturer
hardware costs, and are presented together with total RD&T
costs in Table 3-12. Manufacturer costs are shown to be
$7,671,000 undiscounted and $7,869,000 discounted.
b. Cost to Users
i. First Cost
Manufacturers must recover their costs by increasing the
first-price of vehicles equipped with HDGEs. It is expected
that manufacturers face a 10 percent cost of capital and
recover their RD&T costs in the three model years immediately
following introduction of the standard, 1988-90. The
discounted RD&T costs amortized over the engines projected to
be sold in those three model years results in a cost per engine
of $4.02, or about $4. The sum of the engine share of RD&T
cost and the hardware cost ($3.42) is the first price
increase. Averaged over all model year 1988-90 HDGES, the
total is $7.44, or about $7 per engine.
ii. Operating Costs
As described in the technological feasibility analysis,
the fuel economy impact of the 6.0 NOx standard is expected to
be negligible for HDGEs. Since only engine recalibrations and
component redesigns will be used to achieve the required
emission reductions, maintenance should not be affected by this
standard.
iii. Total User Cost
The total cost to the user is simply the first price
increase of approximately $7 per vehicle equipped with an
HDGE. Operating'costs are not expected to change.
2. Total Manufacturer and. User Costs ' for the 1988
Standard
a. Manufacturer Cost
The total manufacturer cost of compliance for the 1988
HDGE NOx standard of 6.0 g/BHP-hr for the three model years
-------
1983
1989
1990
1991
1992
1993
- - 1994
1995
1996
1997
1998
1999
Pro1 acted
Gas__
389
386
384
381
379
381
383
384
385
388
392
396
3-25
Table 3-11
[ HDE Sales (thousands)
Diesel
338
353
367
382
397
403
409
416
423
429
433
437
Total
727
739
751
763
776
784
792
800
808
817
825
833
Based on information presented in Reference 3.
-------
3-26
Table 3-12
HDGE Manufacturer Costs for 1988 NOx Standard
Undiscounted
RD&T Hardware
1,594K
2,114K
$1,330K
1.320K
1,313K
Discounted*
RD&T
$1.929K
2,325K
-
-
-
Hardware
-
-
$1,330K
1.200K
1,085K
1988
1989
1990
$3,708K $3,963K
Total $7,671K 7,869K
* 10 percent to 1988.
** Research and development costs
*** Certification costs.
-------
3-27
1988-90 is the surr of fixed and variable costs developed above,
and is about 57.7 rr.il lion undisccunted or $7.9 million
discounted at 10 percent to the year of the standard.
b. User Cost
The user cost is the sum of the first price increase
developed above and any change in operating costs due to the
standard. No operating cost increases are expected, so that
the average cost to the user of a model year 1988-90 vehicle
"with a HDGE is about $7.
3. 1991 NOx
a. Cost to Manufacturer
EPA received only one comment concerning the cost estimate
per HDGE due to the proposed 4.0 NOx standard. Chrysler
estimated a cost of $180 for reduction from 6.0 to 4.1
g/BHP-hr, compared to the estimate in the NPRM of $18.
However, Chrysler's comment did not detail the technology which
would cause this price increase, nor did it indicate the amount
of research, overhead and markup contained in the estimate. It
is therefore difficult to determine to what extent the
difference is based on actual differences between EPA and
Chrysler estimates of specific costs, and to what extent it is
due to differences in assumptions in areas such as
technological approach, mark-up, vehicles over which costs are
apportioned, etc.
Nevertheless, absent any detailed comment, EPA has
ra_evaluate<3 the cost of the 1991 HDGE NOx standard based on
the revision of the HDGE NOx portion of the technological
feasibility analysis.
i. Fixed Cost
The cost analysis for the proposed 4.0 HDGE NOx standard
contained in the Draft RIA included RD&T costs for the
recalibration of the fuel, EGR, and ignition systems,
combustion chamber modifications, arid for certification. The
final cost analysis for the 1991 HDGE NOx standard includes
RD&T costs in these areas plus others for improvement of
secondary air management and intake manifold modifications for
those HDGE families not already receiving -these changes. These
costs are allocated as described below. The per family cost to
accomplish each of these tasks is the same as allocated for
1988. For convenience, these costs are shown again in Table
3-13.
-------
3-23
Table 3-13
RDS.T Costs per HDGE Family
Costs per
Tasks Family
1. Fuel, ignition and EGR recalibration $39,600
2. Secondary air management $13,200
3. Combustion chamber redesign $306,900
4. Intake manifold modifications $66,000
5. Certification $192,170
-------
3-29
First, coses are again allocated to each HDGE family for
further fuel, ignition, and EGR recalibration work. However,
this is probably conservative since it is reasonable to expect
that some families will be able to meet the 1991 NOx standard
with only minor changes to 1988-90 calibrations.
Second, further costs for the more significant changes
(secondary air management, combustion chamber redesign, and
intake manifold modifications) are now allocated for these
eight families not receiving these changes in meeting the 1988
standard. This is also conservative, since it is unlikely that
all eight families would require all three of the more
significant changes. Thus, as is shown • in Table 3-14,
recalibration work for all 11 families totals to $435,600 and
other more significant changes for the remaining eight families
totals to $3,088,800.
- Third, certification costs are once again appropriated for
all 11 families at a total cost of $2,113,870. Once again,
this is conservative, since it is likely that some families
will be able to gain 1991 certification through a running
change in lieu of full certification.
As shown in Table 3-14 when the work is allocated as
discussed above and summed, RO&T costs total $5,638,000
(undiscounted).
ii. Variable Cost
As was discussed above, redesigned hardware may be
necessary for the combustion chamber and the intake manifold
modifications, and was conservatively estimated above to cost
$10 per affected engine. For the 5.0 standard, 8 of the 11
HDGE families will require this hardware. Spreading this cost
evenly over all HDGEs, the average cost per engine is $7.27.
iii. Total Manufacturer Cost
The total manufacturer cost, including RD&T and hardware,
sums to $13,933,000 undiscounted and $14,153,000 when
discounted at 10 percent to 1991. This includes the RD&T costs
developed above and the costs for redesigned hardware on model
year 1991-93 engines. The stream of costs, both undiscounted
and discounted, is shown in Table 3-15.
b. Costs to Users
i. First Cost
The incremental increase in the first price of a 1991 HDGV
over a'similar 1990 HDGV can best be presented as the average
-------
3-30
Table 3-14
Sumrr.acv of 1591 HDGE NOx 3D&T Costs
1.
2.
Category
Fuel, ingition and EGR
recalibration
Secondary air management,
Families Cost per
Affected Family To
11
8
$39,600
$386, 100
$435,
$3,088,
tal
600
800
combustion chamber redesign,
and intake manifold mods.
Certification " " 11 $192,170 $2.113,870
$5,638,270
-------
3-31
Table 3-15
HDGE Manufacturer Costs for 1991 NOx Standard
Undiscounted
RD&T
3,524K
2,114K
-
-
••
S5,638K
Hardware
-
-
2,770K
2,755K
2*7 "Jr\ V
, 1 /UK
$8,295K
Discounted*
RD&T Hardware
$4,264K
2,325K
2,770K
2,505K
2.289K
$6,589K $7,564K
1989**
1990**1
1991
1992
1993 -
Total $13,933K $14,153K
*10 percent to 1991.
** Research and development costs,
*** Certification costs.
-------
3-32
first price increase expected if costs are spread ever all
HDGEs. If RD&T costs are amortized over three years of sales
(1991-93) at a 10 percent cost of capital, the average per
engine increase attributable to RD&T equals $6.33. This added
to a fleet average hardware cost of $7.27 gives a short term
average first price increase of $13.60 for the 1991 NOx
standard. In the long term this cost drops to about $7.
ii . Operating Costs
As is discussed in the technological feasibility analysis,
no significant fuel economy impact is expected for HDGEs due to
the 5.0 standard. Therefore, fuel costs will not be affected.
Increased maintenance is not expected as a result of meeting
the 1991 HDGE NOx standard, so maintenance costs should not
change.
iii. Total User Cost
The total user cost of the 1991 standard is simply the
first cost increase, averaging $13.60 over the cost of a model
year -1990 vehicle equipped with an HDGE. No increases in
operating costs are expected.
4. Total Manufacturer and User Cost for the 1991
Standard
a. Manufacturer Cost
The total manufacturer cost of compliance for the 1991
HDGE NOx standard of 5.0 g/BHP-hr for the three model years
1991-93 is the sum of fixed and variable costs developed above,
and is about $13.6 million undiscounted or $14.2 million
discounted at 10 percent to the year of the standard.
b. User Cost
The user cost is the sum of the first price increase
developed above and any change in operating costs due to the
standard. No operating cost increases are expected, so that
the average cost increase to the user of a model year 1991-93
vehicle with a HDGE is $13.60. After RD&T costs are amortized,
the first price increase will drop to about $7 per HDGV.
C. HOPE NOx and Particulate Standards
Specific comments on the proposed HDDE NOx standards were
received from five manufacturers — Cummins, Ford, General
Motors, International Harvester, and Mack, as well as from the
Department of Energy, the Engine Manufacturers Association, and
the American Trucking Association. All commented either that
-------
3-33
hardware costs were substantially higher or fuel economy losses
considerably greater, or both, than those estimated by EPA. In
general, detailed derivations of the costs were not given in
the comments, and EPA's derivation approach was not
challenged. However, in one case a comparison of assumptions
was made which detailed the reasons for fuel economy cost
estimate differences without investigating the relative merit
of the two methods.
The comments have prompted revised analyses of
manufacturer cost estimates by EPA. In the case of RD&T, costs
are based on the number of engine families which will require
the work; at this time, family-specific data- on HDDEs is not
available. Therefore, EPA can only estimate the number of
families which will require RD&T allocations based on general
manufacturer comments. For hardware costs, those components
which were not costed by Rath & Strong[1] were estimated by EPA
in the Draft RIA; again, they are updated here based on
general manufacturer comments. These estimates are retail
price equivalent (RPE) costs. The detailed reanalysis is
provided in the following sections.
1. 1988 NOx Standard
a. Cost to HDDE Manufacturers
Hardware changes deemed necessary for engines to meet the
6.0 g/BHP-hr NOx standard as outlined in the Draft RIA included
injection timing retard and the addition of aftercooling to
non-aftercooled turbocharged engines. In comparison to this,
Ford outlined hardware plans of improved fuel injection
systems, variable injection timing, and turbocharging on all
engines, as well as charge air cooling on some. Cummins listed
variable injection timing, low temperature aftercooling,
increased fuel injection pressure, combustion chamber
modifications, and an electronically controlled fuel system,
while International Harvester listed engine cooling system
•changes, air-to-air aftercooling, and electronically controlled
fuel systems. The dissimilarities in these lists of hardware
changes contributed to the difference in cost estimates between
manufactures and EPA; they also prompted a revision of
development tasks and hardware in the EPA analysis, which is
presented later.
The costs presented by manufacturers for HDDE 6.0 g/BHP-hr
NOx control are compared to EPA's projections as follows:
-------
3-34
Retail Price Increase per HOPE, 6.0 NOx Standard
Draft Analysis $16
$78
Cummins $100-800
HDDE average
HDDE wich new technology
Depending
timing;
control
on need
including
for variable
particulate
Ford
IH
$350
$337
HDDE average,
control
HDDE average
including particulate
-• Ford's cost is an estimate of the consumer cost for the
hardware changes described above, which are currently planned
for model year 1987 in anticipation of more stringent NOx
standards for that year. Cummins indicates "...an increase in
estimated engine prices for the Cummins product line." The
value given by International Harvester (IH) is a sales-weighted
value of going from 10.7 to 6.0 g/BHP-hr NOx based on the cost
difference between present California and Federal IH engines.
The California NOx standard is 5.1 g/BHP-hr.
The costs presented by manufacturers are clearly higher
than those given by EPA, prompting the reanalysis which is
included below; however, since the industry estimates do 'not
give detailed breakdowns of components and costs, it is
difficult to tell whether the values in the comments can be
directly compared to the EPA estimates. For example, the
industry estimates are presumably per engine requiring new
technology, although this is not clearly stated; the EPA
estimates presented in the NPRM are spread over all HDDEs, some
of which are projected not to require the new hardware: Also,
manufacturers, when indicating increase in "consumer cost", do
not indicate which RD&T costs are included, nor the amount of
dealer markup. From aggregate RD&T costs which were provided
confidentially by some manufacturers, it appears that ongoing
basic research costs are included, rather than only those costs
which arise from research directly applicable to this standard,
as in the EPA analysis.
Dealer markup is presumably higher in the industry
analysis than in EPA's analysis; EPA bases its markup on the
idea that the dealer will incur no costs due to the standard
except for the interest that must be paid on the higher cost of
the inventory before it is liquidated; this interest is
included in the EPA markup value. Using this method, the
dealer will receive no profit due to the standard, but will not
take a loss either. On the other hand, if the manufacturers
-------
3-35
are including their usual dealer markup in their "consumer
cost" estimates, the dealer is caking a profit from the
standard; such dealer profit is not correctly applied to the
cost c.f the standard.
Such differences in the analyses may partially explain the
differences between EPA and manufacturer cost estimates, and
create a situation where the values cannot be directly
compared. Resolution of these potential differences is
confounded by the fact that the development of the industry
cost estimates is not documented in the comments, so that even
the discussion above is only a conjecture as to what the cost
values presented in the comments may actually represent.
Nevertheless, as part of its review of the technological
feasibility of the 6.0 g/BHP-hr NOx standard, EPA has
reevaluated the control technology needed to meet the
standard. This in turn has led to a reanalysis and revision of
the cost figures; this revision is discussed below.
i. Fixed Cost
The reanalysis of the RD&T and hardware costs necessary
for HDDEs to comply with the 6.0 g/BHP-hr NOx standard includes
the timing retard and addition of aftercooling as in the Draft
RIA, plus additional RO&T and hardware costs of improved
aftercooling, variable injection timing, and improved
turbocharging. The number of HDOE families remains at 86, as
in the Draft RIA.
Timing retard calibration evaluation was costed at $26,400
per engine family in the Draft RIA, based on three calibrations-
per engine family, 160 labor hours per calibration at a rate of"
$50 per hour, and a 10 percent contingency factor. This value
has been increased to $132,000, five-fold the original, because
an increased number of calibration evaluations would be
necessary to optimize fuel economy and to deal with the larger
number of approaches available for meeting the 6.0 g/BHP-hr
standard. For 1986 engine families, RD&T comes to $11,352,000
for timing retard. The addition of aftercooling to 10 percent
of the HDDE families (half of those turbocharged engines
without aftercooling) remains as before, at $57,400 per engine
family and a total of $494,000 based on six person-months of
engineering and development work per family.
New to this analysis for the 6.0 standard is the
improvement of aftercooling, which was previously believed to
be necessary only for the 4.0 standard but is now added in
response to manufacturer comments. This RD&T cost also remains
the same as in the 4.0 portion of the Draft RIA, at $172,200
per family for air-to-air and $57,400 for air-to-liquid
-------
3-36
aftercooler systems. At a 50 percent application rate for each
system for the 72 engine_ families expected to have
aftercooling, the total cost is $8,266,000.
Variable injection timing (VIT) and improved turbocharging
are also new to this analysis, and are each estimated to have
RD&T costs of $95,700 per engine family; however, half of these
costs are applied to the particulate standard, leaving $47,850
per family for each of the two tasks. This value is based on
two designs per change, four person-months per design, and the
usual $50 per hour and 10 percent contingency factor, as well
as an additional 25 percent to account for the effort needed to
optimize fuel economy. Assuming VIT and improved turbocharging
will each be assigned to 50 percent of the 86 engine families,
these costs are $2,058,000 per tasK, or $4,115,000.
Certification costs remain at $6,500,000 as presented in
the Draft RIA. This includes dynamometer time and emission
test costs for one durability and three data engines per engine
family.
Total RD&T costs are then the sum of all these costs, or
$30,738,000. This is comprised of $11,352,000 for timing
calibration evaluations, $494,000 for the addition of
aftercooling, $8,266,000 for the upgrading of current
aftercooling systems, $4,115,000 for VlT/improved turbocharging
and $6,500,000 for certification.
ii. Variable Cost
Hardware costs per engine have also increased. While
costs for injection timing retard and addition of aftercooling
remain, total hardware costs are increased due to the addition
of. improved aftercooling, variable timing, and improved
turbocharging. The per engine hardware cost for HDDEs adding
aftercooling capability remains at $61, with 10 percent of
KDDEs being affected.
Improved aftercooling cost per engine also remains as
originally in the Draft RIA for the proposed 4.0 standard, at
$73 for conversion of an air-to-liquid to an air-to-air
aftercooler, and $91 for upgrading of an existing air-to-liquid
system. These costs, however, are now being allocated to some
engines which will be built under the 6.0 standard, in response
to manufacturer comments that some will need the technology for
the earlier standard. The rate of application is such that
half of all turbocharged engines (31 percent of all HDDEs) will
employ new or upgraded aftercooler systems for the 1988
standard. One-third of this 31 percent, or 10 percent of the
total, is comprised of HDDEs getting air-to-liquid
-------
3-37
aftercooling for the first time as described above. Of the
remaining 21 percent, 16 percent will convert to improved
air-to-liquid aftercooling and 5 percent will convert to
air-to-air aftercooling. These costs average to $87 for each
vehicle converting to improved aftercooling, and $18 when
applied to all HDDEs.
The incremental cost of electronically-controlled variable
injection timing is estimated at $25 per engine, and is applied
to half of the engines, with half of the cost charged to the
particulate standard. As noted above, this is an EPA estimate
based on manufacturer comments to the NPRM.
Improved turbocharging is estimated to cost $5 per engine
as in the Draft RIA, and would apply to 50 percent of all
turbocharged HDDEs (31 percent of all HDDEs). Half of this
cost would be applied to the particulate standard.
The sum of these costs on a fleetwide average basis is
about $32 per HDDE. When applied only to engines requiring the
new technology, and the average hardware cost for the 6.0
g/BHP-hr NOx standard would be $93 per engine.
iii. Total Manufacturer Cost
Total RD&T and hardware costs must be discounted at 10
percent to the year of the standard, 1988, in order to
represent actual manufacturer cost. It is reasonable to expect
that RD&T expenditures will be made in the two years
immediately preceding the year of the new standard. The RD&T
costs described above are summed in Table 3-16, and presented
in undiscounted and discounted form. Hardware costs are
allocated according to sales projections, which have been
updated due to new information[3] and were shown in Table
3-11. Using these sales projections and the average cost per
HDDE developed above results in the distribution of hardware
costs shown in Table 3-17, in both undiscounted and discounted
form. Since the 6.0 g/BHP-hr NOx would only apply through
1990, the hardware costs are presented and summed for only
three model years of HDDE sales. Manufacturer hardware costs
sum to $34 million dollars undiscounted and $31 million
discounted.
These manufacturer costs for RD&T and hardware are
summarized and presented on an annual basis in Table 3-18.
This analysis results in a total undiscounted manufacturer cost
of about $64.4 million and a discounted cost of about $66.4
million.
-------
3-38
Table 3-16
HDDE RDS.T Co'sts got 6.0 NQx
Undiscounted Discounted*
Non- Non-
Cert. Costs Cert. Costs Total Cert. Costs Cert. Costs Total
1986 $17,OOOK .Sl.OOOK S18.000K $20,570K $1,210K S21.780K
1987 $ 7,200K $5,50QK SI2.7QOK $7,920K $6,050K $13,97QK
TOTALS -$24,200K $6,500K $30.700K $28,490K $7,260K $35,750K
Discounted at 10 percent to 1988.
-------
3-39
Table 3-17
HDDE Hardware Costs for 1983 NQx Standard
1988
1989
1990
TOTAL
Undiscounted
$10,753K
11.228K
11,675K
S33.656K
Discounted*
$10,753K
10.207K
9.649K
. S30.609K
Discounted at 10 percent to 1988.
-------
3-40
Table 3-18
HDDE Manufacturer Costs for 1988 NOx Standard
1986
1987
1938
1989
1990
TOTAL . .
Und is conn ted
RD&T Hardware
$ia,oooK
12.700K
$10,753K
11.228K
11,675K
$30.700K $33,656K
$64,356K
Di scounted*
RD&T
$21,780K
$13,970K
—
—
__
$35,750K
$66.,
Hardwa re
--
—
$10,753K
" ' 10,207K
9.649K
$30,609K
359K
Discounted at 10 percent to 1988.
-------
3-41
b. Cost to Users
i. First Cost
Increases in HDDE purchase price due to the 6.0 NOx
standard are determined in the same manner as for the LDT
standard, except that the capital costs (RD&T) are expected to
be recovered in three rather than five model years due to the
introduction of the second NOx standard in 1991. The average
increase in first cost of HDDEs would consist of the sum of the
discounted RD&T cost amortized over vehicle sales for model
years 1988 through 1990 through plus the average per engine
hardware cost. These costs can also be expressed per HDDE
requiring new technology rather than as average per engine cost
by adding the RD&T cost apportioned only over the affected
vehicles to the cost of the hardware required. Costs using
these two different approaches are presented below.
First, total discounted RD&T cost amortized over the total
HDDE sales projected for model years 1988 through 1990 results
in a cost of approximately $37 per vehicle. When this is added
to the $32 average hardware cost developed above, the average
first price increase is $69. When distributed only over those
engines affected by the standard, the costs are $50 for RD&T
and $93 for hardware, for a total of $143 for a vehicle
receiving new technology.
ii. Fuel Economy
In Chapter 2, Technological Feasibility, it was estimated
that fuel economy penalties associated with the 1988 model year
NOx standard would be in the tange c£ C to 2 percent in the
short term. This penalty should tend to disappear by the time
of implementation of the second standard, 1991, as part of a
normal trend toward further engine and vehicle improvements.
EPA has reevaluated the cost impact of these short-term fuel
economy losses based on the comments received. This is
presented below.
First, fuel economy estimates for 1988 HDDVs have been
updated, and are derived from information in Reference 3.
These estimates for LHDDEs, MHDDEs, and HHDDEs have been
lowered to 15.1 mpg, 8.0 mpg, and 5.9 mpg respectively. This
makes them closer to the Argonne National Laboratory estimates
supplied in comments received from the Department of Energy
(DOE) which compared EPA and ANL assumptions. These fuel
economy values are combined with a fuel cost of $1.20 per
gallon; ANL used $1.46 as the estimated average cost over the
lifetime of the vehicle. However, since the price of diesel
fuel has varied significantly in the recent past, and since
fuel prices continue to be highly sensitive to the world
-------
3-42
political climate and, therefore, unpredictable, EPA has used
S1.20/gal as representative of ^today's price without attempting
to project future price increases.
Average annual mileages and lifetimes remain as in the
Draft RIA, at 11,000 miles per year for 10 years, 30,000 miles
per year for 9 years, and 65,000 miles per year for 8 years for
LHDDEs, MHDDEs, and HHDDEs, respectively. These values include
one rebuild for some of the MHDDEs and most of the HHDDEs, and
are reasonable estimates of the actual lifetimes of these
engines for fuel - economy purposes. The useful life VMTs used
by ANL do not involve any rebuilds, but EPA has found that the
majority of the heavier HDDEs are not, in fact, retired after
their initial useful life, and hence would continue to accrue
fuel economy penalties. Therefore, EPA has included these
higher lifetime values in calculating the lifetime fuel economy
cost for the standard.
A 10 percent discount rate is employed with the values
given above and the fuel economy estimates are sales weighted
in arriving at the average cost per engine and total average
lifetime cost. The sales fractions used are 35 percent LHDDE,
29 percent MHDDE, and 36 percent HHDDE, and are derived from
information presented in Reference 3.
Using the fuel economy, fuel price, and vehicle/engine
average lifetime miles and years it can be calculated that e.ach
one percent reduction in fuel economy corresponds to an annual
increase in diesel fuel usage of 7.3 gallons for LHDDEs, 37.5
gallons for MHDDEs, and 110.2 gallons for HHDDEs. These
increases in fuel usage correspond to lifetime discounted costs
Of $54 for LHDDEs, $259 for MHDDEs, and $705 for HHDDEs. Sales
weighting these costs gives the average lifetime cost for a 1
percent change in fuel economy of $348 per affected engine.
Applying these average costs, the range in the fuel
economy cost per engine which corresponds to the 0 to 2 percent
change expected for model year 1988 vehicles for the 1988 HDDE
NOx standard is $0 to $696." This value should drop to $0
before implementation of the 1991 standard.
iii. Maintenance
No increase or decrease in maintenance is expected as a
result of the application of the technology needed to meet the
6.0 'g/BHP-hr NOx standard and hence there should be no impact
on costs.
-------
3-43
iv. Total User Costs
In summary, owners of model year 1988 through 1990 HDDVs
can be expected to pay an average of approximately $69 more for
the emission control on these vehicles than they would have
paid without promulgation of the NOx standard. In terms of
fuel costs, the increased average lifetime cost per vehicle is
expected to be between $0 and $696, tapering off to $0 in later
model years. Total lifetime increase is thus $69 to $765 in
the short term, and $69 in the long term.
2. 1988 Particulate Standard
a. Cost to HOPE Manufacturers
i. Fixed Cost
The RD&T costs for the non-trap particulate standard were
reevaluated, but due to the lack of specific comments, EPA saw
no need for major change from those costs presented in the
NPRM. Some revisions in the 1988 particulate RD&T costs are
caused by changes in the RD&T costs for 1988 NOx control which
are allocated equally with particulate control, and general
comments indicating the need for more development to deter fuel
economy penalties.
The original non-certification RD&T cost was based on four
tasks: 1) modifications to the combustion chamber through
changes in the piston, 2) changes in injectors and increased
injection pressure, 3) changes in the fuel delivery system to
refine air/fuel ratio control during transient operation, and
4) changes to the turbocharger to improve air delivery
characteristics during transient operation. In the Draft RIA,
the cost per task to accomplish this non certification RD&T was
estimated at $3,292,000. This was based on 2 design
evaluations per task, 4 person-months per design at $50 per
hour, and a 10 percent contingency factor, applied to one-half
of the 86 engine families. EPA determined that one-half of the
families would need the work based on manufacturer comments to
the NPRM.
The current estimate for RD&T is based on the same four
tasks listed above, as well as on one-half of the cost of
applying variable injection timing (VIT). The other half of
the cost for developing VIT is included in the RD&T costs for
NOx, as is half of the cost of improved turbocharging. Thus,
the particulate standard is being allotted the full RD&T .cost
for three tasks — piston modification, transient air/fuel
ratio control, and improvement in injectors — as well as half
the cost for each of the two tasks of improved turbocharging
and VIT. The cost per task in the present analysis is
increased from the
-------
3-44
original estimate by 25 percent in order to account for an
additional effort to optimize, fuel economy, in response to
comments that such an effort will occur. The estimated
non-certification RD&T cost for the 1988 particulate standard
is therefore $3,292,000 per task X (3 + 2(1/2)) tasks X 1.25
fuel economy effort factor or about $16.5 million.
In the Draft RIA, 1988 HDDE certification was estimated to
cost $13 million. Assigning this cost at 50 percent each for
NOx and particulate allots $6.5 million of the certification
.costs to particulate control. This brings the total
undiscounted RD&T cost to approximately $23 million.
ii. Variable Cost
Hardware costs, like fixed costs, are calculated much as
in the Draft RIA, where they were estimated at $20 per affected
engine, based on $5 per modified component. The modified
•--components include: 1) combustion chamber/piston design
changes, 2) injector and injection pressure modifications, 3)
fuel delivery system changes, and 4) turbocharger
improvements. This analysis has changed only to reflect the
changes discussed above.
The cost remains at $5 per component for the first three
components listed above, while the charge for improved
turbocharging is halved to $2.50, with the other $2.50 being
allotted to the cost of the 1988 NOx standard. In response to
, limited comments in this area, an additional $5 is included for
improvements in transient control of air/fuel ratio control and
• turbocharger operation. As was mentioned in the cost analysis
-" for the 1988 NOx standard, electronically controlled variable
injection timing will be used to control NOx and particulate.
Adding this capability is expected to cost $25 per engine.
Half of the cost of variable injection timing, or $12.50 per
engine, is also now charged to the particulate standard to
accompany such a charge being added to the NOx standard.
Summing these costs results in a $35 cost for each engine
receiving the modifications; overall, half are expected to
receive them. The average cost per HDDE is therefore about $18.
iii. Total Manufacturer Cost
It is expected that the RD&T costs for the 1988
particulate standard will be incurred according to the time
table shown in Table 3-19, which is proportional to that
presented in the Draft RIA. Costs are shown both undiscounted
and discounted to 1988 at 10 percent. Total hardware costs for
the three model years following introduction of the standard
are based on projected sales for those years as shown in Table
3-11; these costs are estimated using projected sales figures
and are given in Table 3-20 in both undiscounted and discounted
forms.
-------
3-45
Table 3-19
- HDDS RD&T Costs for 1988 Particulate
Undiscounted Discounted*
Non- Non-~
Certification Certification Total Certification Certification Total
1986 $15,OOOK $1,OOOK $16,OOOK $18,150K $1,210K $19,360K
1987 $1,500K $5,500K $7,OOOK $1,650K ' $6,050K $7,700K
TOTAL $1S,500K $6,500K $23,OOOK 4l9,800K $7,260K $27,060K
Discounted at 10 percent to 1988.
-------
3-45
Table 3-20
HDDE Hardware CJSts "Eoc 1933 Pacticula-a
1988
1989
1990
TOTALS
L'ndiscounted
$5,915K
S6,178K
S6.422K
$18.515K
Discounted*
$5,915K
$5,616K
$5.307K
$16,838K
Discounted to 10 percent to 1988.
-------
3-47
Total manufacturer cost is the sum of these total RD&T and
hardware costs, and amounts co approximately $41.5 million
undiscounted and $43.9 million discounted cost, as shown in
Table 3-21.
b. Cost to Users
i. First Cost
The total RD&T cost developed above can be recovered by
increasing HDDE prices by $28 for model year 1988-90 engines.
When added to the average hardware cost of $18, the total first
price increase averages $46 per HDDE. Apportioning this cost
only over those vehicles affected by the standard results in a
first price increase of about $84 per HDDE.
ii . Operating Cost
As described in the technological feasibility analysis,
neither fuel economy nor maintenance is expected to be impacted
by the 0.6 g/BHP-hr particulate standard, and hence will not
impact user costs.
iii. Total User Cost
The average increase in user cost due to the 1988
particulate standard is the sum of the first price increase and
any increase in operating costs. Operating costs are not
expected to change, so the average user cost is simply the
first cost increase of $46 per model year 1988-90 vehicle
equipped with an HDDE. «•
3. Total Manufacturer and User Costs for 1988 NOx and
Particulate Standards
The total HDDE manufacturer cost of compliance with the
1988 standards is developed above for the NOx and particulate
standards separately. These costs are shown together in Table
3-22, and total approximately $110 million manufacturer cost
discounted to 1988.
The total HDDE user cost per vehicle is also developed
above separately for the two standards; the total is shown in
Table 3-23 and is $115-$810, depending on the fuel economy
penalty. This value will tend towards $115 in later model
years as fuel economy improves.
-------
3-43
Taole 3-21
HDDE Manufacturer Costs for 1983 "articulate
1986
1987
1988
1989 -•
1990
Totals
Undi
RD&T
$16,OOOK
$7,QQOK
--
__
$23,OOOK
$4
scounted
Hardware
--
—
$5,915K
$6,178K
$6,422K
$18,515K
1.515K
Discounted*
RD&T Hardware
$19,360K
$7,700K
$5,915K
$5,616K
$5,307K
$27,060K $16,838K
$43,393K
Discounted to 10 percent to 1988.
-------
3-49
Table 3-22
Total HDDE Manufacturer Costs
1988 MOx and Particulate Standards
Undi
RD&T
30,700K
23,OOOK
53,700K
scounted
Hardware**
$33,656K
L8,515K
52.171K
Discounted*
RD&T
$35,750K
27,060K
62,810K
Hardware**
$30,609K
16,838K
47,447K
NOx
Particulatc
Total
Grand Total $105.871K
$110.257K
* Discounted at 10 percent to 1988
** ' Model year 1988-90 HDDVs.
-------
3-50 .
Table 3-23
Total HDDE User Costs
1988 NOx and Particulate Standards
Fleetwide Vehicle Average
First Cost Fuel Economy
NOx " $ 69 $0-696
Particulate $ 46 $0
Total 4115 $0-696
Grand Total • $115-810
-The $696 fuel economy cost is from a short-term 2 percent
fuel economy penalty.
-------
3-51
1991 NOx
a.
Cost to HDDE Manufacturers
The manufacturer comments which applied to the 6.0 HDDE
NOx standard generally applied to the originally proposed 4.0
standard also, with the basic assertion that EPA cost estimates
were too low. Specific values for the manufacturer costs of
meeting the lower standard were given only by Ford and the
Department of Energy (DOE), as follows:
Retail Price
Draft
Analysis
Increase Per HDDE, 4.
$291
$347
from
from
6
6
.0
.0
to
to
.0
4
4
NOx Standard
.0,
.0,
HDDE
HDDE
average
with new
technolog
Ford
DOE
$350 from 6.0 to "lowest possible NOx"
$700 from 10.7 to "lowest possible NOx"
$643 from 10.7 to 4.0
As in the analysis for the 1988 standard, it is unclear
what is included in the cost estimates presented in the
comments in regard to such things as RD&T, markup, and percent
of engines over which costs are apportioned. The technology
changes which are being used to estimate these costs are also
not detailed in the comments, although Ford states that all
engine models will require air-to-air aftercooling. And
finally, DOE presents its value, developed by Argonne National
Laboratories (ANL), as the total cost of going from the current
NOx level to the final proposed level, rather than as the
incremental costs involved with the intermediate level, as EPA
does. Ford presents costs for both the total and incremental
reductions, but finds the costs to achieve the total reduction
without any discounting of fixed costs; DOE also does not
discount, making it difficult to directly compare with EPA's
estimates.
However, the
those projected by
technology and using
standard. The larger
include the total cost
approximately twice as high
presumably due to the cost of
costs given in the comments are close to
EPA, when apportioned over engines with new
incremental cost from the intermediate
values given by Ford and DOE which
of controlling from 10.7 to 4.0 NOx are
as EPA's incremental estimate,
the intermediate standard, which
was addressed above.
Therefore, the analysis of HDDE manufacturer costs for the
5.0 NOx standard remains essentially the same as that in the
Draft RIA for the 4.0 standard, with changes only in hardware
costs in order to reflect comments and to complement changes
made for the intermediate standard.
-------
3-52
i. Fixed Cost
RD&T costs for the proposed 4.0 g/BHP-hr NOx standard were
developed in the Draft RIA. The RD&T costs for a 5.0 g/BHP-hr
NOx standard are essentially the same as for the proposed
standard at $28,700,000 undiscounted cost, but are delayed one
year, along with the standard.
ii. Variable Cost
Hardware costs applicable to the 5.0 NOx standard for
HDDEs accrue* from additional and improved aftercooling, piston
design and turbocharging. The Draft RIA also included costs
to cover some portion of the costs for applying electronic
control modules (ECMs). However,-since manufacturers' comments
have indicated that virtually all engines will have such units
for reasons other than emission reductions prior to
implementation of the standard, costs for electronic control
modules are not properly attributable to this standard. Thus,
the total hardware cost estimates are reduced from those in the
Draft RIA. The other component costs remain the same, however,
and are based on comparisons to costs of similar pieces of
equipment on existing engines.
Based on manufacturers' comments, engines applying
aftercooling for the first time are most likely to .use
air-to-air aftercooling. As discussed in the Draft RIA, the
application of air-to-air aftercooling is estimated to cost
$134. It was shown in the Draft RIA that 21 percent of all
HDDEs are turbocharged and employ no aftercooling; 10 percent
were allocated funds for applying aftercooling in response to
the 1988 standard, leaving 11 percent still without
aftercooling. Applying the $134 per engine to this 11 percent
results in an average of $15 per HDDE. The turbocharged
engines which did not receive new or improved aftercooling for
the intermediate standard will require it now, at a cost of $73
for converting from air-to-liquid and $91 for upgrading an
existing air-to-liquid system. These costs are taken from the
Draft RIA and are the same as used above for the 6.0 standard.
For the 6.0 NOx standard it was projected that 5 percent of all
HDDEs would convert from air-to-liquid to air-to-air
aftercooling and 16 percent would upgrade current air-to-liquid
systems. For the 1991 standard it is projected that another 5
percent will convert to air-to-air and 15 percent will upgrade
air-to-liquid systems. On a weighted basis, the average cost
of improved aftercooling is $17 per HDDE.
The component cost and amount of application of piston
redesign remains as in the Draft RIA, at $5 per engine at a 25
percent application rate. EPA's best estimate based on
manufacturer comments results in $1.25 per HDDE.
-------
3-53
Earlier in this analysis, it was projected that 50 percent
of the HDDEs would need turbocharger improvements to ree- tne
1988 NOx and particulate standards. Improved turbocharging is
now expected to be employed on the remaining half of the
engine's to meet the 1991 NOx standard at a cost of $5 per
engine, as allotted for the intermediate standard. This
results in $2.50 per HDDE.
EGR is eliminated as an expenditure to meet the 1991 NOx
standard, in a response to indications from the manufacturers
that they will not employ EGR on their engines. Electronic
control module hardware costs are also eliminated, as discussed
above, although RD&T costs were allocated for software design.
Total hardware costs are then the total of the above costs
for aftercooling, piston design, and improved turbocharging.
This amounts to $36 per engine average hardware cost, or $113
per HDDE receiving the new hardware.
iii. Total Manufacturer Cost
To calculate total manufacturer cost, RD&T and hardware
costs must be discounted to the year of the standard, 1991.
The distribution of RD&T costs which was given in the Draft RIA
and is used again here is shown in Table 3-24. Hardware costs
expended according to sales projections and discounted to the
year of the standard are shown in Table 3-25. The total
manufacturer cost arises from the sum of these costs, and is
developed in undiscounted and discounted forms in Table 3-26.
Total manufacturer cost of compliance with the 1991 HDDE NOx
standard is shown to be about $71 million undiscounted and $73
million discounted tc 1991
b. Cost to Users
i. First Cost
Incremental increases in first cost due to the 5.0 HDDE
NOx standard are determined in the same manner as described
previously, except fixed costs are recovered over 1991 through
1993 model year HDDEs. The average 'increase in first cost of
HDDEs would consist of the sum of the discounted RD&T costs
amortized over sales for model year 1991 through 1993 plus the
hardware cost developed above. These costs are approximately
$32 for RD&T and $36 for hardware for a total of $68 average
HDDE first price increase. These "costs can also be expressed
per HDDE requiring new technology rather than as average per
engine cost by adding RD&T cost apportioned only over these
engines to the cost of the hardware required. These costs are
approximately $44 for RD&T and $113 for hardware, for a total
for $157 for an engine receiving all of the new technology.
-------
3-54
Table 3-24
HDDE aD&T Costs for 199L NOx
Undiscour.ted
Non-
1988
1989
1990
Totals
Cert.
$7,
14,
' 1.
$22,
Costs Cert. Costs Total
OOOK
OOOK - 1,OOOK
200K - 5.500K
200K $6,500K
$7,
15,
6,
$28,
OOOK
OOOK
700K
700K
Cert
$9
16
1
$27
Discounted*
Non-
. Costs
,317K
,940K
,320K
,577K
Cert. Costs Total
$9,
1,210K 18,
6,050K ' 7,
$7,260K $34,
317K
150K
37QK
837K
Discounted at 10 percent to 1991.
-------
3-55
Tacle 3-25
HDDE Hardware Costs for 1991 NOx
Sales
1991
1992
1993
TOTAL
Undiscounted
$13,673K
14,209K
14f 423K
42,305K
Discounted
$13,673K
12,917K
11,920K
38.510K
Sales taken from Table 3-11.
** Discounted at 10 percent to 1991.
-------
3-55
Table 3-26
1988
1939
1990
. -_ -lg91
._1992
--. - 1993
Undiscounted
RD&T Hardware
$7,OOOK
15,OOOK
- 6,700K
- $13,673K
14,2Q9K
14.423K
S28.700K $42,305K
Discounted*
RD&T Hardware
$9,317K
S18.150K
7,370K
$13,673K
12,917K
11.920K
$34,837K $38,510K
TOTAL
$71,005K
$73,347K
Discounted at 10 percent to 1991.
-------
3-57
The cumulative increase in first cost over current engines
to achieve compliance with the 1991 standard would be the SU.TI
of the costs for 1988 and 1991. The total increase in first
cost i.s thus $69 for 1988 hardware plus $68 for 1991 or $137.
ii. Fuel Economy
In Chapter 2, Technological Feasibility, it was estimated
that fuel economy penalties associated with the 1991 standard
would be in the range of 0 to 1 percent in the short term, and
this penalty should tend to decrease to one-half percent with
time as vehicle and engine improvements are made. The 0 to 2
percent penalty associated with the 1988 standard should have
•disappeared by 1991. - - -
Fleetwide fuel economy costs are calculated in the same
manner as for the 1988 standard, amounting to an average per
vehicle lifetime increase of $348 per 1 percent increase in
fuel consumption. With a 0 to 1 percent change in fuel economy
expected for the 5.0 standard, the short term fuel cost
increase is thus $0 to $348, tapering off to $0 to $174 over
the long term.
Several commenters—Mack, American Trucking Association,
and Department of Energy—indicated the fuel cost increases
would be much greater than this. All estimates, however, were
based on a variety of different assumptions regarding vehicle
lifetimes and amount of fuel currently used, as well as on
higher fuel economy penalties. The fuel economy penalty issue
is most important, and is addressed in Chapter 2, Technological
Feasibility; the other issues are methodology differences of
VMT and fuel price estimates, and are discussed above in regard
to the 1988 HDDE NOx standard.
i i i. Maintenance
Maintenance is not expected to be affected by this
•standard, and hence should not impact on cost.
iv. Total User Cost
In summary, owners of model year 1991 and later vehicles
which are equipped with HDDEs can be expected to pay an average
of approximately $68 incrementally over model year 1988 through
1990 vehicle prices or $137 total more than they would have
paid without the introduction of "the two HDDE NOx standards.
In terms of fuel costs, the increased average lifetime cost per
vehicle is expected to be between $0 and $348, tapering off in
later model years to $0 to $174. Incremental lifetime increase
is thus $68 to $416 in the short term, and $68 to $242 in the
long term. Total lifetime increase for model year 1991 and
-------
3-58
later HDDEs due to the NOx standards is $137 to $485 in the
short term, and $137 to $311 in the long term. These costs are
summarized in Table 3-27.
5. 1991 Diesel Particulate Standards (0.25 g/BHP-hr for
HDDEs with 0.10 g/BHP-hr for Urban Buses)
In this section, the costs of the 1991 diesel particulate
standards for HDOEs are examined. As described in the
Technical Feasibility Chapter (Chapter 2), achieving these
standards will require the use of trap-oxidizer technology on
100 percent of the urban buses and about 60-70 percent of .the
remaining HDDEs. Since the same basic type of trap-oxidizer
system will be used on both HDDEs and urban buses, in the
subsequent analysis of comments and cost derivations, the
primary discussion in each section centers on HDDEs in general,
and is then followed by a discussion of any special
considerations of urban buses, as necessary.
• a. - Cost to Manufacturers — ~. _ . ._. .
i. Fixed Cost
In the draft analysis,- EPA separated research and
development costs into three categories: 1) general system
development; 2) specific engine family design;- 3) electronic
control development. The seven largest HDDE manufacturers were
each allotted about $2.8 million to develop general trap
systems. Smaller manufacturers were expected to rely on
guidance from trap-oxidizer manufacturers for general designs.
Engine family specific designs were assumed to be required by
about 70 percent "of the engine families with averaging, at a
cost of about $230,000 per engine family. The development of
electronic controls was estimated at about $115,000 for each
engine family.
Three comments were received regarding EPA's research and
development cost estimates. In the first comment, GM stated
that EPA had clearly underestimated the cost of basic trap
development, claiming that it had already expended $40 million
by the end of 1984. In presenting it's $40 million
expenditure, GM failed to distinguish what portion of this
amount is attributable to LDD trap development and which is
attributable to HDDE trap development. Without this
information, it is impossible to know how much GM has indeed
spent on trap systems for HDDEs. The company's claim can be
placed in perspective, however, by the fact that LDD
regulations requiring trap technology were promulgated for the
1985 model year in early 1980.[4] (These requirements were
subsequently delayed in early 1984 until the 1987 model
year.) [5] By contrast, HDDE trap requirements for the 1991
model year are just now being promulgated in this rulemaking.
-------
3-59
Table 3-27
Heavy Duty Diesel Engines
Discounted* User Cost set Engine for I9S8 and 1991 NOx
SHORT TERM:
RD&T
Hardware
Fuel
TOTAL
1988 Standard
$37
$32
$0 to 696
$69 to 765
1991 Standard
$32
$36
$0 to 348
$68 to 416
Total
$69
$68
$0 to 348
$137 to 485
LONG TERM:
RD&T
Hardware
Fuel
TOTAL
$37
$32
$ 0
$69
$32
$36
$0 to $174
$68 to 242
$69
$68
$0 to 174 •
$137 to 311
At 10 percent per year to the year that standard becomes
effective.
-------
3-60
Additionally, most trap-related technical information submitted
to EPA by GM has concerned light-duty traps. It seerr.s
reasonable to expect, therefore, that only a small portion of
GM's total trap development expenditures should be attributed
to the HDDE trap standards. The Agency would also like to
point out that due to the problematic nature of estimating
development expenditures for each manufacturer, EPA's
projection should be reviewed in terms of an overall average
per manufacturer, with some spending more and others less.
Because of its size, it is not unreasonable to expect that GM
should fall into the former category. Hence, GM's comment
provides no basis for revising EPA's original estimate of
general development costs.
The second comment, which was confidential, identified
another company's expenditures for HDDE trap development from
1979 through 1984. In this case, the reported values were well
within EPA's estimate for each manufacturer. Therefore, the
original estimate for general system development appears to be
appropriate based on this comment.
The third comment came from International Harvester which
claimed that its mechanical durability testing would require 11
LHDDVs and 13 MHDDVs with each vehicle successfully traveling
10&,000 miles and 150,000 miles, respectively. While not
agreeing with the need for such a large test fleet, EPA
calculates the cost of such a program at about $2.3 million.
This amount is less than half of the development cost for IH' as
derived in the draft analysis. Therefore, the comment provides
no basis for changing EPA's original research and development
.estimates. ,
Another area of comment concerning development costs was
that of vehicle modifications. The draft analysis did not
contain a cost for development and tooling expenditures which
might be needed to modify the vehicle assembly to integrate the
trap-oxidizer into the overall design. International Harvester
expressed the strongest concerns regarding the potential
magnitude of vehicle modifications. In an apparent reference
to. tractor-trailer combinations, IH stated that if two traps
are necessary and their size requires that they be mounted
behind the cab, then the location of such things as the sleeper
unit, fuel tank, air tanks, fuel and oil filters, aerodynamic
side shields, etc. may be affected. This would in turn
adversely affect hundreds of body builders. General Motors
stated that in its evaluation of possible vehicle design
changes, there appears to be adequate room within the vehicle
frame on a MHDDV to mount a trap and muffler, although a few
vehicle components may need to be relocated on some versions.
For its HHDVs, GM claimed that if mounted behind the cab, traps
may restrict the vehicles turning radius which, in turn, may
reduce tractor-trailer offerings. General Motors also alledged
-------
3-51
that "essentially" no space was available for a trap in it's
urban bus, and that significant redesign of this vehicle would
be required. The company claimed that potential changes may
include relocating air conditioner and heater components,
eliminating seating for up to five passengers, or installing
new suspension systems and bulkheads. Cummins generally
commented on the need for vehicle modifications without
identifying specific changes required.
EPA agrees with the commenters to the extent that vehicle
modifications may be required for certain trucks in order to
accommodate trap oxidizers. The extent of any required vehicle
modifications will obviously be dictated by -the type of trap
system ultimately chosen by manufacturers, and upon its
specific configuration. Since neither the final trap type nor
specific configurations have yet been identified, the potential
costs associated with required vehicle redesign can not be
-.quantified to any degree. However, it is possible to discuss
in general terms the types of trucks that are most likely to be
affected and how the negative effects of any redesign might be
minimized.
As stated by GM, the greatest potential for vehicle
modifications is associated with HHDDVs and urban buses.
Regarding HHDDVs, and all other diesel trucks for that matter,
it should be remembered that diesel particulate emissions
averaging will result in a significant number of trucks not
needing traps. To the extent that a manufacturer can
anticipate problem installations, such vehicles/engines might
be excluded from having traps. Beyond this, it is reasonable
to expect that many HHDDVs and urban buses will normally
• undergo some design changes by the 1991 effective date of the
standards, especially in light of the emphasis being placed on
improved aerodynamics by HDDV manufacturers. For such
vehicles, the incremental cost of incorporating traps in the
design would be minimal. Finally, there are many vehicles for
which creative packaging of the trap system will avoid costly
redesigns. For example, Southwest Research Institute is
currently testing a GM coach engine with a trap configured to
replace the engine's exhaust manifold.[6] Such a design would
require no redesign of the urban bus.
Overall, then, while modifications will likely be needed
on some vehicles, the extend of these changes will in large
part be dictated by the engine manufacturers choice of trap
system, and the foresight with which it is configured or
packaged to meet the requirements of the vehicle manufacturer
or body builder. Therefore, EPA believes that the number of
significant design changes can be minimized, and when averaged
over the fleet their impact will be small. Because of this, no
fixed cost for vehicle modifications will be included in the
cost of the regulations.
-------
3-62
In addition to the expense of research and development,
the draft analysis included, the fixed cost of emission
certification testing. No comments were received on this cost
component, so it is being retained here without change.
In summary, none of the comments supported any changes to
the fixed costs of the draft analysis. They are therefore
being retained unchanged. Nonetheless, it is interesting to
note that even if the comments had provided a basis for
revising the fixed cost estimates, any corresponding change in
the total cost of the regulations would be very small. As will
become evident later in this analysis, fixed costs are only
about eight percent of the total cost.
- As shown in Table 3-28, the total fixed cost of the 1991
particulate standards is $49.5 million. The distribution of
the expenditures over time is the same as originally estimated
in the draft analysis, except each allocation has been delayed
one year to account for the revised effective date of the
standards from 1990 to 1991.
ii. Variable Costs
The draft analysis explored the potential use of two
significantly different types of trap oxidizers* for HDDEs: a
non-catalyzed/ ceramic monolith system; and a catalyzed
wire-mesh system. As fully described in the Diesel Particul-ate
Study (DPS),[2] which accompanied the draft analysis, both
systems function similarly in that particulate matter is
.filtered from the exhaust and then periodically burned in the
trap to prevent excessive exhaust backpressures which would
•degrade engine performance and fuel economy. This latter step
is termed "regeneration" and is significantly different in the
two trap types, depending on the presence or absence of a
catalyst. The ceramic monolith design assumed in the draft
analysis used a fuel burner to heat the trapped material to its
ignition point. During this process, the engine exhaust flow
is temporarily routed around the trap, while the burner and
trap are supplied with a controlled air supply to ensure
adequate oxidation of the trapped material without excessive
heating. The regeneration system with a catalyzed wire-mesh
trap can be less complex, since the requisite temperature
increase of the trap is significantly less. The catalyst trap
evaluated in the draft analysis was assumed to achieve ' the
required moderate heat rise by delayed in-cylinder fuel
injection; thereby, eliminating the need for a fuel burner and
bypass system.
In assessing the variable or hardware cost of the two
systems, the draft analysis found that the wire-mesh design
with its catalyst coating was quite expensive. Hence, the
-------
3-63
TaoLe 3-28
Total Fixed Costs o£ the
199 L HDDS Particulate Standards
vear Develooment
L w O fc
1987
1988
1989
- 1990
Total
$8
20
13
_ _2
$43
. OM
.OM
.OM
.OM
.OM
Certification Tot
$8
20
Sl.OM 14
5.5M- _7
S6.5M $49
ai
. OM
.OM
.OM
.SM
.5M
-------
3-64
ceramic monolith trap with a fuel burner regeneration system
was used to estimate the cost of the proposed HDDE particulate
standards. A summary of the hardware costs that were used in
the proposal are shown in Table 3-29.
Comments on the cost of trap-oxidizer systems were
received from six manufacturers, one of which was
confidential. Cost figures were also provided by the
Department of Transportation and the Department of Energy.
Unfortunately, each of the government estimates were supplied
as a composite' total of the trap-oxidizer standards in
conjunction with either the 1988 non-trap standard or the 1991
HDDE NOx standard and, therefore, could not be analyzed in
detail. Hence, these latter two comments are not discussed
further. The non-confidential industry estimates are displayed
in Table 3-30.
The manufacturers estimates in Table 3-30 were generally
reported as the cost of a total system. Little, if any
information was provided as to the derivation or basis of the
estimates. For example, the manufacturers usually provided no
breakdown of the total system into its various cost
components. Also, the estimates were variously described as
"cost to the consumer" or "consumer effect." Therefore, it is
unclear if some of these costs include fixed or operating
costs, or whether they inappropriately reflect the full retail
cost of a trap system, rather than the incremental cost fo-r a
new vehicle. Even if only a 1 percent penalty were included in
some of these estimates, this could add about $350 to the total
cost for an average HDDE when discounted to the year of vehicle
purchase.
Cummins was the only manufacturer shown in Table 3-30 that
provided a breakdown of its system cost by hardware component.
As reported by the company, the component costs are as follows:
1. Trap Substrate Material - $720-$!,080
2. Trap Casing and Ceramic Mounting $250
3. Diesel Burner for Regeneration $400
4. Electric Air Blower for Burner $175
5. Miscellaneous Control Costs $650
Total $2,195-$2,555
Cummins describes the trap substrate, i.e., ceramic
monolith, as being 60-90 liters for a "possible dual trap
option" at $12 per liter. Additionally, the costs are
described as component costs from suppliers, without an
allowance for assembly, machining of other ancillary parts, or
fixed manufacturing costs.
-------
3--J5
Tacle 3-29
Trao-Oxidizer Va
Cost Category
Ceramic Monolith and Hous
Burner
Fuel Delivery System
Fuel Ignition System
Auxiliary Air System
Exhaust Diversion System
Sensors
ECU
Total Hardware
ri=ble Cost
LHDDE
ing $207
7
9
26
30
45
12
37
$363
Fror, the Draf1:
iMHDDE
$343
14
18
21
30
69
24
37
$556
31 A
HHDDE
W^^MM^HM^B
$402
14
18
21
30
105
24
37
$652
-------
3-66
Table 3-~30
Manufacturers' Trap-Oxidizer Cost Esrimat.es
Source Cose Description
IH $1285-2070 MHDDE, single trap
4710 HHDDE, single trap
7210 HHDDE, dual trap if required
Ford 2200 No comment
Cunmins 2810-3270 HHDDE, possible dual trap option
GA 575-900 LHDDE
2300 MHDDE, single trap
- - - 4000 HHDDE, dual trap
Saab 2500+ No comment
-------
3-67
The confidential comment also provided some breakdown of
cost oy component. Due to the confidentiality of the comment,
however, all that can be said is that in contrast to Cummins
estimate, the reported trap cost, (i.e., substrate and housing)
is substantially less expensive, while the air supply and
burner are somewhat more expensive.
Cummins', makes a rigorous analysis of the estimates
impossible. The only clear conclusion that can be reached is
that the cost of a trap-oxidizer system as reflected by the
manufacturers comments is substantially higher than EPA's
estimates from the Draft Analysis (Table 3-29). In order to
address this disparity, EPA's only recourse has been to
completely reevaluate its estimates, using the comments where
possible, to better define the variable costs of the
regulations. In preparing this reevaluation, EPA has also made
-use of an independent contractor to prepare component cost
estimates.
Overview of the Analysis
In reevaluating the variable cost of trap-oxidizers for
HDDEs, EPA will examine the component costs of three separate
systems. The first system uses a ceramic monolith trap in
conjunction with a fuel burner and exhaust bypass for
regeneration. This is very similar to the trap-oxidizer unit
used in the draft analysis to estimate the costs of the
proposal. The second system is the same as the first, in that
a ceramic monolith is used to filter the exhaust, but differs
with regard to the type of heat source used to initiate
regeneration. In this system, electric heating elements are
included in "the trap housing and are energized with the
vehicle's electrical system. An exhaust bypass is also required
with this system. Electrically regenerating the trap can be
advantageous since it eliminates the bulk and safety
considerations associated with -the fuel burner approach. The
third system is radically different from the others in both
trap design and method of regeneration. The filter medium in
this trap design is composed of ceramic fibers which are wound
into a type of fabric. Regenerating the ceramic trap is
accomplished through the use of a metallic catalyst compound
that is injected into the exhaust at the time regeneration is.
desired. Since the catalyst substantially reduces the ignition
temperature of the trap particulate material, no exhaust bypass
system is required. This system is attractive primarily
because its simplicity could result in reduced> costs compared
to the other two systems. Each of these systems will be
further described below.
-------
3-68
Most research and testing to date have focused on the
ceramic monolith trap. Of the two regeneration methods
described above, the fuel burner concept has been successfully
used in vehicle tests and its hardware components are well
understood. The electrical regeneration system, on the other
hand, is much less well defined at the present time. The
ceramic fiber trap and catalytic regeneration system is the
most recent entrant in the trap-oxidizer field. The trap
concept is proprietary to Mercedes-Benz AG and relatively
little is known about it compared to the ceramic monolith
trap. Due to the present state of knowledge, the cose of a
ceramic monolith/fuel burner system can be estimated with .the
greatest degree of certainty. Therefore, as in the draft
analysis, this system will be used to derive the variable costs
associated with the 1991 particulate standards. Also,
considering its state of development, this trap-oxidizer design
.could be the first commercially available system.
The variable costs of the ceramic -monolith/electric system
and ceramic fiber/catalyst system are still of interest,
however, since their potential advantages may result in either
or both of these supplanting the ceramic monolith/fuel burner
design. Therefore, these systems are examined here to provide
a measure of sensitivity to the overall cost estimates.
The variable cost of each trap-oxidizer system is found by
determining the retail price equivalent (RPE) of each component
part. With few exceptions, the costs were based on work
performed by Mueller Associates under contract to EPA.[7, 8]
The contractor's estimates were based on the manufacturer's
cost of each component. To obtain the required RPE of the
"various components, EPA adjusted the contractor's estimates to
reflect the added costs associated with a manufacturer's
overhead and profit, in addition to dealer costs. The mark-up
factor used to derive the RPE of each component was 1.29 (i.e.,
a 29 percent increase) . This factor has been used in past
rulemaking actions and is derived in Reference 9.
The costs not taken directly from the contractor were
estimated by EPA and will be specifically identified were
applicable in the discussion. The Agency's estimates are based
on previous work by Lindgren,[10] information supplied by
Mueller Associates,[7,8] the DPS, and on engineering
evaluations of similar automotive components.
Now that the general methodology has been described,
estimates of the various component costs can be presented.
This will be done first for the two trap designs and then for
the three regeneration systems. After the component parts have
been estimated, tne resulting total cost of each system will be
presented.
-------
3-69
Ceramic Monolith Trap Costs — The cost of any specific
trap design is dependent on the volumetric requirements of the
filtering medium. One of the most important considerations in
sizing, a trap is to make it large enough to avoid undue exhaust
backpressures, which would degrade engine performance and
adversely affect fuel economy. In general, the requisite trap
volume for a given engine is dependent on the volumetric
exhaust flow during normal operation. The HDDE trap sizes
assumed in the proposal were derived in the DPS. The analysis
was based on successful testing of a 5.0 liter trap on a
Mercedes-Benz 300D by Southwest Research Institute for
EPA.[11] Since volumetric exhaust flow from an engine is
roughly a function of the amount of fuel- consumed (i.e.,
inverse of fuel economy), HDDE trap sizes were estimated by
increasing the LDDV trap size (i.e., 5.0 liter) by the ratio of
the vehicle's MPG (26 MPG) and the projected average MPG's of
each HDDE size category. The resulting HDDE trap volumes,
which were used to estimate the cost of the proposal, varied
from 8.3 to 18.4 liters depending on HDDE size.
No comments were received on this method of sizing HDDE
, traps. The Agency's estimated trap volumes, however, contrast
sharply with cost comment from Cummins indicating the use of a
60-90 liter trap in its calculations. Unfortunately, Cummins
provided no information describing how it arrived at this
size. In addition, trap test data supplied by GM (described
further below) are based upon trap volumes somewhat greater
,v than estimated by EPA. This disparity has caused EPA to
,, reevaluate its sizing methodology.
The major difficulty in attempting to estimate exhaust
flow changes from one vehicle or engine type tc another is in
choosing a sizing parameter that accurately reflects the many
variables which ultimately determine the actual exhaust
volume. Such key variables include air/fuel ratio, vehicle and
engine speed, engine efficiencies, and how the engine is loaded
in normal operation, i.e., what percentage of the engine's
rated horsepower is typically used. As vehicles become more
disparate in size and function, the accuracy of any single
parameter for estimating exhaust volumes will diminish due to
the multitude of variable interactions. This notwithstanding,
EPA continues to believe that fuel consumption is a reasonable
surrogate for approximating exhaust flows. It inherently
accounts for many of the changes in vehicle operating regimes
and engine operating parameters among vehicle types.
While fuel consumption appears to be a useful method, for
estimating exhaust flows, EPA also recognizes that the wide
disparity in operating regimes of LDDV engines and some HDDEs,
especially the heaviest trucks, could result in this approach
being somewhat inaccurate for such HDDEs. Therefore, at least
for some applications, engine horsepower also may be a useful
sizing parameter.
-------
3-70
To explore this method, a review of EPA trap-oxidizer test
programs and the information submitted in response to the.
proposal was conducted. Thi's review showed that several
different combinations of trap volume and engine horsepower
have been tested. Two EPA test programs are useful in this
comparison. First, the previously referenced Mercedes-Benz
300D test was conducted with a 5.0 liter ceramic monolith
trap. The rated power of the engine used in this vehicle model
is 118 HP. [12] Second, EPA has also tested a GM bus engine
with a 20 liter trap. [13] This MHDDE has a rated power of 190
HP.[12] While -the results were quite variable, overall
essentially no fuel economy penalty was observed with this trap
. volume/engine size combination. The trap volume (liters) to
horsepower ratios from these tests are about 0.04 for the
"EPA-LDDV" and about 0.10 for the "EPA-MHDDE."
The comments contained test data for two additional
engines. The first engine has a rating of 205 HP, [12] and was
tested by GM with trap sizes ranging from 20-25 liters. The
information submitted by GM for some of these tests shows what
appears to be quite reasonable pressure drops across the trap
during actual over-the-road vehicle testing. Hence, this trap
volume/engine combination may represent an acceptable trap
volume to horsepower ratio from the standpoint of minimizing
any potential fuel economy penalty. The average trap volume to
horsepower ratio for the "GM-MHDDE" is 0.11. The comments also
. contain confidential test results on a LHDDE vehicle. Due to
the confidential nature of the comment, however, all that 'can
be stated is that the trap volume to horsepower ratio for this
test was somewhat less than the EPA-MHDDE factor. The Cummins
comment regarding trap volumes is not used here since the basis
of the estimate was not reported.
Table 3-31 presents a summary of the average trap volumes
that result from applying the various factors discussed above
(i.e., both MPG and horsepower based) to the average fuel
economy and horsepower ratings for the various HDDEs. Note
that the MPG values shown in the table have been updated from
those in the DPS, as discussed in the section on the 1991 HDDE
NOx standard. This has resulted in revised HDDE trap volumes
using the fuel consumption sizing method.
From the table, it is readily apparent that the various
approaches result in a wide range of estimates. The lowest
values are consistently estimated by the EPA-LDDV horsepower
approach. This is not surprising given that a trap would
likely be sized or optimized for the most typical type of
operation. The power requirements of a LDDV under normal
operation is usually less than that for a diesel truck when
expressed as a percentage of the engine's rated horsepower.
For LHDDEs the difference may be rather small, but the
-------
3-71
3-31
of HPEE Trap Volume Esrimacea
Fuel Economy and Horsepower Based Volumes
THTTW 15 1 130 8.6L
« K 'S SS
5.2L
7.4L
14. OL
13. OL
18. 5L
35. OL
14. 4L
20.4L
38.5L
-------
3-72
disparity becomes progressively greater as the HDDE size is
increased. As a result, the EPA-LDDV horsepower factor would
progressively underestimate HDDE trap volume requires as the
HDDE size of the vehicle increased.
At this point, some judgment must be used in conjunction
with the remaining fuel economy based and horsepower based
values shown in Table 3-31 to identify reasonable HDDE trap
volumes. An extremely important consideration in this decision
is the tradeoff between trap volume and effects on fuel
economy. In this respect, it can generally be concluded that
the cost of a somewhat larger trap is much less than the c.ost
of adversely affecting fuel economy by using a trap that is too
small. Hence, it is EPA's intent to be conservative, (i.e., to
err . toward larger volumes) in estimating trap volume
requirements.
Most LHDDEs are loaded and driven much like LDDVs. This
argues strongly in favor of using the fuel economy based
estimate, since this method should be quite accurate in this
case. To be conservative, however, the trap size for this
category will be estimated by averaging the EPA LDDV fuel
economy based volume with the various horsepower based
volumes. The result is an estimated volume of 11 liters for
LHDDEs. Considering that the operating regimes-of MHDDEs and
HHDDEs become increasingly dissimilar to LDDVs and LHDDEs as
truck size grows, and that the EPA and GM horsepower factors
are based on tests that should not result in undue fuel economy
penalties, the GM-MHDDE horsepower factor will be used to
estimate trap volumes for these vehicles. The resulting
estimates are 21 liters for MHDDE and 39 liters for HHDDEs.
Another important detail which must be dealt with before
the trap costs can be estimated is the number of traps that may
be required by the variously sized HDDEs. The "DPS assumed the
number of traps for each HDDE size as follows: one for LHDDEs,
two for MHDDEs, and two for HHDDEs. No comments were received
regarding the number of traps for LHDDEs. The comments from IH
and GM indicate one trap will be sufficient for MHDDEs (Table
3-30). GM provided an illustration of this concept that showed
two ceramic monoliths arranged in series to provide the
necessary volume.
The comments for HHDDEs are indecisive with regard to- the
number of traps per vehicle (Table 3-30) . The Cummins and IH
comments suggest single traps are possible. Also the Cummins
comment regarding the possible dual trap requirement would seem
to be invalidated given that the trap volume requirements
estimated above are significantly less than assumed in its
description. GM's comment was provided in the context of its
12.1 liter, 435 HP engine. This engine is significantly larger
-------
3-73
than post of the engines in this HDDE size category. It is
possible such a large engine -night require two traps. However,
even in this situation, placing trap monoliths in series to
attain, the required volume is possible, although some increase
in bacKpressure would result. Based on the comments and the
requisite trap volumes, EPA believes that for most if not all
HDDEs, a single trap is sufficient if not preferable for system
simplicity.
The ceramic monolith trap has several components. These
components include the monolith itself, a ceramic mat and a
trap housing. Each component and its associated cost is
discussed below.
The ceramic monolith material is constructed as a matrix
of alternatively opened and closed cells. Particulate material
is collected as the exhaust flows through the porous wall of
one channel into the next. The monoliths used for HDDEs are
assumed to be approximately 12 inches in diameter, although
smaller sizes can also be made. The cost is estimated at about
$6 per liter, based on information from Corning, one of the
largest ceramic monolith manufacturers.[8] This can be
contrasted with the $12 per liter cited by Cummins in its
comment, which was unreferenced. Using the former value, the
estimated price of the ceramic monolith material for the
various HDDEs is: $66 for LHDDEs, $126 for MHDDEs, and $234
for HHDDEs.
The ceramic mat holds the monolith securely within the
housing. It also functions as a shock absorber and provides
thermal insulation. This item is estimated at $3, $6, and $12
for LHDDEs. MHDDEs, and HHDDEs, respectively.[8]
The trap housing encloses the ceramic monolith and ceramic
mat. It includes baffles, flanges, and pipe connectors (used
to connect the trap to the exhaust system and fuel burner or
bypass valve), in addition to fittings for mounting sensors.
The estimated cost is $32 for LHDDEs, $40 for MHDDEs, and $46
for HHDDEs.[8]
Table 3-32 presents the total estimated cost of a ceramic
monolith trap for each of the HDDE size categories.
Ceramic Fiber Trap Costs — Other than the basic
construction of this trap design, little specific information
is available. In general, perforated stainless steel cylinders
are wound with silica fibers until the desired filter "fabric"
has been created. Several such tubes are then arranged in
parallel inside the stainless steel housing so that the exhaust
must pass through the fabric and into the stainless cylinder
before ' exiting the trap. To estimate the cost of this trap
design, the volumetric requirements that were developed for the
-------
3-74
Table 3-32
P^Mmated HD-DE Trap Costs
Tr.n Design . LHDDE _MHDDE_ _HHDDE.
Ceramic Monolith $101 $1" $292
Ceramic Fiber 73 106
-------
3-75
ceramic monolith are assurred co apply. The cost for the entire
cera-ic trap is estinated at 573 for LHDDE, $106 for KKD3E, and
$140 for HHDDE.[7] These costs are also shown in Table 3-32.
Fuel Burner Regeneration System Costs -- A typical system
of this type has several primary components: a burner can, a
fuel delivery system, an ignition system, an auxiliary air
supply system, an exhaust diversion system, and an electronic
control system. An additional component required by the
trap-oxidizer which is actually neither part of the trap nor
regeneration system is the exhaust pipe. The cost of this
component and the others are discussed below.
The burner can is located just upstream of the trap. The
can is designed to contain the flame and distribute the heat
output (e.g., about 100,000 Btu/hour) evenly across the face of
the trap. Additionally, the unit provides a location for
mounting the fuel injection nozzle, ignition plug and flame
sensor, and auxiliary air injection nozzle. Due to the
operating environment and required long life, the burner can is
largely made of high grade stainless steel. The basic cost of
this component is relatively insensitive to variations in heat
output requirements. Therefore, the estimated cost of $21 is
used for all HDDEs.[7]
The fuel delivery system is composed of a fuel injector,
control solenoid, fuel line, and fuel line connectors. This
system is used in conjunction with the vehicle's existing fuel
injection system. The use of an electric solenoid provides
precise control of the regeneration rate, and provides
effective overtemperature protection. The fuel injector and
solenoid is estimated at $12 for all HDDEs.[7] The fuel line
and connectors are estimated by EPA to cost about $2 per
vehicle.
The ignition system provides spark ignition and flame
control. The components include an electrode, an inverter, and
a step-up voltage transformer for generating a high-voltage
discharge. Also, the system includes a flame sensor and sensor
relay as a safety consideration for cutting off fuel to the
burner if combustion fails. Mueller Associates estimated the
cost of a continuous spark system.[8] The Agency finds that a
somewhat smaller system providing a periodic spark, when used
with the flame sensor, is fully satisfactory. As a result,- EPA
estimates the system to be approximately 20 percent less than
the'contractor's estimate, or $35 for all HDDEs.
The auxiliary air system supplies a controlled amount of
air to the burner and trap during regeneration. Its components
include an air pump, a control valve operated by an electric
-------
3-76
solenoid, and an air delivery line with connectors. The air
pump is criven by the vehicle engine and is a larger, more
durable version of those already used on light trucks and
heavy-duty vehicles. It is equipped with a check valve to
prevenc exhaust backflow into the air pump. Since air is only
required during regeneration, the air pump is assumed to be
equipped with an electric clutch so that it can be disengaged
from the engine in order to save fuel. The Agency estimates
the cost of the air pump and electric clutch to be about $45
and the associated air delivery tubing with connectors at about
$5 for all HDDEs. The control valve and electric solenoid for
this system are estimated at $14 per vehicle.[7]
The exhaust diversion system consists primarily of a
bypass valve and a solenoid controlled actuator that
temporarily reroutes the exhaust around the trap during
regeneration. The bypass valve is a butterfly type constructed
of stainless steel, located just upstream of the combustor. In
estimating the price of this unit, the cost of a stainless
steel exhaust pipe has been included. This pipe will replace
the standard steel exhaust pipe that normally extends from the
engine manifold, or turbocharger, to the muffler. In the DPS,
the cost of the stainless steel pipe included a credit for the
standard steel pipe which it replaced. In this analysis, the
standard steel pipe is assumed to be roughly equivalent to that
required to bypass the trap. Therefore, the full cost of a
stainless steel exhaust pipe is included in the cost of .the
bypass valve. This component is estimated at $49 for LHDDEs,
$52 for MHDDEs, and $58 for HHDDEs.[8] The electric
solenoid/actuator is estimated to cost $15 for all HDDEs.[8]
To initiate and control regeneration, several different
sensors, an electronic control unit, and wiring harness will be
required. A backpressure sensor will detect the need for
regeneration and is estimated to cost $17 per HDDE.[7] A trap
temperature sensor estimated to cost $5, and will be used to
protect the trap from excessive heat.[8] A sensor will also be
used to ensure the engine has reached the proper temperature
before regeneration is initiated. This sensor was estimated to
cost about $1 in the DPS.
Regarding the electronic control unit, manufacturers are
expected to equip essentially all HDDEs with such units by the
1990s, irrespective of emission standards. For this reason,
the electronic capability required for trap regeneration will
be added to the existing unit." This incremental cost is
estimated at about $34 per HDDE.[8] The wiring harness to
connect the sensors to the electronic control unit is estimated
to cost $14 for LHDDEs and $18 for MHDDEs and HDDEs.[8]
-------
3-77
Based on the above discussion, the total cost of the fuel
eurner regeneration syste.T is estimated to be S269 for LHDDEs,
$276 for MHDDES, and $282 for HHDDEs. The various component
costs are summarized in Table 3-33.
Electric Regeneration System Costs — This type of
regeneration system uses many of the same component parts that
are required by the fuel burner regeneration system. These
include an auxiliary air supply system, exhaust diversion
system, and electronic control system. The exception is, of
course, the replacement of the fuel burner with an electrical
heating system. The costs for each of these systems will be
estimated below, with the discussion focusing on the components
that are different from those described for the fuel burner
system.
The auxiliary air supply uses an air pump and electric
clutch, as required by the fuel burner system, except that the
pump is somewhat smaller in size because air is provided only
to the trap. The Agency estimates this to cost about $40 per
HDDE. The control valve and solenoid is retained at $14, as is
the air delivery line and connectors at $2 per vehicle.
The exhaust diversion and electronic control systems
remain unchanged from those used in conjunction with the fuel
burner. The exhaust diversion system was estimated at $64 for
LHDDEs, $67 for MHDDEs, and $73 for HHDDEs. The electronic
control system was estimated at $71 for LHDDEs and $75 for the
other HDDE size categories.
The cost of the electrical heating system is difficult to
estimate because the specific requirements of the system are
yet to be well defined. The Agency envisions a rather modest
system that depends on the vehicle's existing electrical
system. In this concept, the additional electrical power
required for regeneration is provided by using the existing
batteries in conjunction with a larger alternator. The
electric current from the battery is conducted by cable to the
electric resistance heating elements which are mounted in the
trap housing. The cable is equipped with a fusable link to
protect the batteries and charging equipment in the event of a
short circuit. The power supply to the electric heating
element is controlled by the electronic control unit through
the use of an electromechanical relay. Based on alternator
costs supplied by Mueller Associates, EPA estimates the
incremental cost of the requisite alternator to be about $47
for LHDDEs, $56 for MHDDEs, and $67 for HHDDEs. The cable with
a fusable link is estimated to cost $7, while the relay is
estimated at $10 per vehicle. The electrical heating elements
with mounting hardware are estimated to cost $14 for LHDDEs and
$18 for the other HDDE size categories.
-------
Table 3-33
HOPE Costs for Trap Regeneration System
Cost Category
Burner Can
Fuel Delivery System
Fuel Ignition System
Auxiliary Air System ;
Lxhaust Diversion System '
Electronic Control System
Electrical System
Catalyst Dispensor System
Catalyst
Catalyst System Exhaust Mods
Fuel Burner
UtDDE
$21
14
35
64
64
71
-
Ml [DDE
$21
14
35
64
67
75
-
IOIDOE
$21
14
. 35
64
73
75
-
Electrical
LI DDE
M
-
-
$56
64
71
78
KHDDE
—
-
-
$56
67
75
89
HIIDDE
^
-
-
$56
73
75
102
Catalyst
UfflDE
^
-
-
-
-
$59
-
Ml (DDE
••
• -
-
-
-
$62
-
HIIDDC
mf
-
-
-
-
$62
-
Total
$269
$276
53
5
33
53
9
42
$282
$269 $287
$306
53
18
71
$150 $166 $204
CJ
•vj
00
-------
3-79
EPA's estimated electrical heating system costs range from
S7S-S102, depending on HDDE size. Tr.e Agency believes thec-3
costs are representative of the type of electrical systems that
are commercially viable in the 1990s. However, due to the
current lack of information regarding specific designs, these
cost estimates are subject to a significant degree of
uncertainty. In discussions with various suppliers of
electrical heating equipment, Mueller Associates identified
several important uncertainties that could significantly affect
the cost of the electrical system. These include the size and
power requirements of the heating elements, ensuring adequate
battery capacity for engine starting, and prevention of trap
thermal stress due to uneven heating during regeneration. In
response, Mueller Associates has estimated the potential cost
of sophisticated electrical systems that address each of the
potential areas of concern as suggested by its industry
contacts. Such electrical systems could cost nearly four times
more than that estimated by EPA. The Agency believes that such
sophisticated systems will be found to be unnecessary as more
information becomes available. Hence, only EPA's estimate will
be used in this analysis.
As summarized in Table 3-33, electric regeneration systems
are estimated to cost about $269, $287, and $306 for LHDOEs,
MHDDEs, and HHDDEs, respectively.
Catalytic Regeneration System Costs - Unlike the other
regeneration systems, which are at least conceptually well
known, the catalyst and the technique which will be used to
introduce it into the trap remains a matter of some
conjecture. The approach assumed in this analysis involves
onboard vehicle storage of a metallic compound in dry powder
form. From time to time, a metered amount of catalyst is
fluidized by compressed air from the vehicle's turbocharger and
then injected into the exhaust stream just ahead of the trap to
initiate combustion. This type of regeneration system is
potentially the simplest with regard to the required hardware.
The primary components consist of the catalyst, the catalyst
dispensor system, and the electronic control system. As
discussed in conjunction with the other regeneration systems,
the requisite stainless steel exhaust pipe is treated as a part
of the catalyst regeneration system.
The metallic catalyst is assumed to be copper in the form
of powdered copper chloride (CuCl) . The amount of catalyst
required for new HDDEs is estimated to be about 2.3 pounds for
LHDDEs, 4.0 pounds for MHDDEs, and 8.5 pounds for HHDDEs. This
is based on the allowable maintenance intervals for each HDDE,
the estimated fuel economies for these vehicles as described
previously, and an assumed catalyst requirement equivalent to
0.16 g/gallon of diesel fuel.[14] Combining this with an
-------
3-80
estimated cost of 51.46/pound of CuCl,[8] the catalyst costs
are estimated to be $5, $9, and $13 for LHDDEs, MHDDEs, and
HHDDEs, respectively.
Tne catalyst dispenser system has several parts. A one
gallon polyethylene bottle is used to hold the required powered
catalyst. This is estimated to cost about $1 per vehicle.[7]
The reservoir will be attached to a metering device with an
integral agitator extending into the reservoir to break up
lumps of catalytic material. This device is estimated to cost
about $30 for each HDDE.[7] The metering device uses a small
electric motor with an estimated cost of $12 per vehicle.[7]
After being metered, the catalyst is fluidized and sprayed or
injected into the exhaust using high pressure air. The
fluidizer/injector unit includes the required fittings
necessary to mate it to the metering device and air supply.
The estimated cost is $7 per HDDE.[7]
The final components of this system are the hoses for
transmitting the compressed air and fluidized catalyst. The
Agency estimates the cost of these items at about $3 per
vehicle.
The electronic control system requires four sensors.
Three of these are the same as used by the other types of
regeneration systems: trap temperature $5, engine temperature
$1, and exhaust backpressure $17. An engine speed sensor, is
also used and has an estimated cost of about $5 per MODE. [8]
The electronic control unit requirements are again incremental
to the existing capability, although the cost is somewhat less
'than for the other regeneration systems because the catalyst
system is less complex. The incremental electronic control
unit cost is estimated by EPA to be about $22 for all HDDEs.
The wiring harness will also be less costly. This item is
estimated by EPA at about $9 for LHDDEs and $12 for MHDDE and
HHDDEs.
The standard steel exhaust pipe will be replaced with a
stainless steel counterpart as with the other regeneration
systems. However, the catalyst regeneration system should not
require that the exhaust bypass the trap during the
regeneration process. Therefore, the cost of the stainless
steel pipe should include a credit for the deleted standard
steel pipe. The incremental exhaust pipe cost for each HDDE
size category is taken from the corresponding estimate used in
the proposal, with one revision. As described in the DPS, the
incremental exhaust pipe cost was based on the assumption .that
about 25 percent of the MHDDEs and HHDDEs would have dual
exhausts. This assumption has been revised because essentially
all MHDDEs and HHDDEs are expected to be equipped with
turbochargers in the 1990's and, therefore, will likely have
-------
3-81
only one exhaust pipe. Making this revision, the estimated
incremental cost of the stainless steel exhaust pipe is 533 for
LHDDEs, $42 for MHDDEs, and $71 for HHDDEs.
As summarized in Table 3-33, catalyst regeneration systems
are estimated to cost $150 for LHDDEs, $166 for MHDDEs, and
$204 for HHDDEs.
Total Trap-Oxidizer Variable Costs — Table 3-34 presents
a summary of the three trap system costs. As shown, the
ceramic monolith trap with a fuel burner regeneration system
has an estimated cost of about $370 per LHDDE, $448 for MHD.DE,
and $574 per HHDDE. The ceramic monolith trap with an electric
regeneration system has the potential of costing about the same
as the ceramic monolith/fuel burner system. The ceramic fiber
trap with a catalyst regeneration system could prove to be the
least expensive trap-oxidizer with estimated costs of $223,
$272, and $344 for LHDDEs, MHDDEs, and HHDDEs, respectively.
As stated previously, the ceramic monolith/fuel burner
trap-oxidizer will be used to determine the costs of the
particulate standard, due to the greater uncertainty associated
with the other designs.
Now that the variable costs for each HDDE size category
have been determined, the average hardware cost for each trap
equipped non-bus HDDE and urban bus, as well as that for the
fleet-average vehicle can be calculated. This is done by
combining information on sales and trap usage with the system
cost for the appropriate HDDE class or classes. The
methodology for deriving the various average costs is described
below. This methodology will be used in this section for
variable costs and in subsequent sections as required.
Identifying the variable cost for a trap-equipped urban
bus is the most straight forward. Due to their horsepower
ratings, all urban bus engines generally can be classified as
MHDDEs. In addition, all urban bus engines will require a trap
oxidizer to achieve the 0.1 standard. Therefore, the variable
cost for the average urban bus is simply the value identified
for MHDDEs, or $448 (Table 3-34).
The average cost for a non-bus HDDE with a trap oxidizer
is found by sales weighting the system cost for each size
category. As described elsewhere, total HDDE sales are
composed of 35 percent LHDDEs, 29 percent MHDDEs, and 36
percent HHDDEs. However, bus sales must be removed from this
distribution to find the percentage of sales in each category
for non-bus HDDEs only. Urban bus sales are quite volatile
from year to year, but would generally not exceed about 2
percent of total HDDE sales. Using this percentage and the
fact that all urban bus engines are MHDDEs, the non-bus HDDE
-------
3-82
•Sable 3-J4
HDDS Costs for Trap-Oxidizer Systems
Ceramic Monolith/ Ceramic Monolith/
Size
Category
LKDCE
4HDCE
BJDCE
Fuel Burner Electrical Ceramic Fiber/Catalyst
Trap
101
172
292
System
269
276
282
Total
370
448
574
Trap
101
172
292
System
269
287
306
Total
370
459
598
Trap
73
106
140
System
150
166
204
TotaJ
223
272
344
-------
3-83
sales account for 98 percent of the total and are distributed
as follows: 36 percent for LHDDEs, 27 percent MHDDEs, and 37
percent HHDDEs. Therefore, sales weighting the various system
costs with this distribution results in a variable cost for the
average affected non-bus HDDE of $467.
The per vehicle cost when averaged over the entire fleet
is the sum of the sales-weighted cost for an urban bus and the
sales-weighted cost for the average non-bus HDDE (i.e., both
trapped and untrapped). The urban bus component is simply 2
percent of the cost figure for a MHDDE. The non-bus component
is 98 percent of the average non-bus cost. This average, is
determined by combining the various HDDE category costs with
the non-bus sales distribution and the percentage of the
non-bus fleet that is trap equipped. Expressed generically in
equation form, the non-bus component of the fleet-average
vehicle cost is:
Non-Bus Portion of Fleet-Average Cost »
(.98) X [$ per LHDDE x LHDDE Fraction) + ($ per MHDDE x MHDDE
Fraction) + ($ per HHDDE x HHDDE Fraction)] x (Trap Fraction)
The last term in this equation adds some complexity to the
calculation since the number of traps required for non-bus
HDDEs is projected to change from about 70 percent in 1991, to
about 60 percent in 1994 if the 0.25 were retained in that
year. For ease of presentation, the non-bus cost component of
the fleet-average vehicle will be evaluated as a short term
value representing 70 percent traps and a long term value
representing 60 percent traps. When these non-bus components
are combined with the urban bus component, the result is the
cost of a fleet-average vehicle in the short-term (1991) and
the long-term (1994). The fleet-average cost for intervening
years can be linearly interpolated. Using this methodology,
the variable cost of the 1991 standards when averaged over the
entire HDDE fleet is $329 in the short term and $284 in the
long term.
iv. Total Manufacturers Cost of the 1991 Particulate
Standards
The total undiscounted and discounted costs to
manufacturers are shown in Table 3-35. The fixed costs are
reproduced from Table 3-28. The variable costs are the
products of the hardware cost per fleet-average vehicle and
HDDE sales as shown in Table 3-10. The total undiscounted cost
is $420.9 million, while the discounted cost is $402.6 million.
-------
3-84
Tfcble 3-35
HDDE Manufacturers' Cost
for the 1991 Particulate Standards
Undisoounted Discounted*
"fear RD&T Variable Cost Total Total
1987 $8.01 - $8.01
1988 20. CM - 20.CM
1989 14.CM - - - 14.CM
1990 7.94 - 7.34
1991 . .: - $126.CM 126.CM
1992 - 125.31 125.3^
1993 - 119.9M 119.94
$420.9M
Discounted at 10 percent 10 1991.
-------
3-85
b. Costs to Users for HDDEs Complying with the 1991
Particulate Standards.
i. First Cost
The amount that manufacturers must increase the price of
each HDDE depends principally on the variable cost per vehicle,
the number of vehicles over which the fixed costs will be
apportioned, and on the cost of capital to the manufacturer.
For the 1991 standards, it is assumed that manufacturers will
generally recover their fixed costs prior to the effective date
of the more stringent 1994 HDDE particulate standard. It. is
further assumed that the cost of capital is 10 percent. Hence,
the first cost increase for a vehicle is the sum of a portion
of the discounted fixed cost .and the hardware cost, as
described earlier.
. ._. In the short term the purchase price increment for
trap-equipped non-bus HDDE is estimated at $457 for LHDDEs,
$535 for MHDDEs, and $661 for HHDDEs. This averages $553 for a
trap-equipped non-bus HDDE. The first price for an urban bus
is $535. Expressed on a fleet-average basis, the cost would be
about $390. In the long term, assuming a manufacturer
continues to charge the same per vehicle for its fixed cost
recovery, the fleet-average vehicle would cost an additional
$336.
ii. Fuel Economy
In the draft economic impact analysis, non-bus HDDEs with
traps were estimated to incur a 1 to 2 percent fuel penalty.
Urban buses were assumed tc incur a 2 percent penalty. As
discussed in the Technical Feasibility Chapter, EPA's estimates
generally fell within the range of fuel penalties that were
presented in the few comments on this issue. Also as discussed
in that chapter, the trap volumes in this final analysis have
been significantly enlarged from those assumed in the
proposal. As a result, the upper range of EPA's previous
estimate has been revised downward from a 2 percent penalty to
a -1.5 percent penalty; Therefore, the new range for non-bus
HDDEs equipped with traps is 1.0-1.5 percent, while the new
point estimate -for urban buses is 1.5 percent.
Several important methodological changes have been made in
calculating the fuel economy impact for each affected vehicle.
For non-bus HDDEs, the estimated MPG values for each HDDE size
category has been revised to reflect updated estimates. This
was fully described in the previous discussion of the 1991 HDDE
NOx standard where the discounted lifetime cost of a 1 percent
penalty for each size category was shown to be the following:
$54 LHDHEs, $259 MHDDEs, and $705 HHDDEs. Using
-------
3-36
this information, the 1.0-1.5 percent penalty represents a
discounted Lifetime cost of $35_0-$525 for the average non-ous
HDDE equipped with a trap.
For urban buses, three key assumptions have been revised
from the values used in the draft analysis. A recent EPA
analysis shows the average annual bus mileage is about 45,000
miles rather than 50,000 miles, and that the average lifetime
is closer to 12 years than 10 years. [3] These two changes are
included in this updated analysis. The third change affects
fuel costs for buses. The draft analysis utilized the same
cost per gallon of diesel fuel for both non-bus and bus HDDEs,
i.e., $1.20/gallon. Diesel fuel for urban buses is actually
significantly less costly than that for non-bus HDDEs due to
.volume discounts and lower taxes. A comment from the
Department of Transportation supported the use of a
$1.00/gallon cost for urban buses. This value has been adopted
for use in this analysis. Based on these revised values, the
.estimated 1.5 percent fuel penalty results in a discounted
lifetime fuel cost of $1070 for each urban bus.' Expressed on a
fleetwide basis, the average HDDE will incur a short-term fuel
economy penalty of about $261-$381 and a long-term penalty of
about $227-$330.
It should be noted that the above fuel economy penalties
were estimated for trap-oxidizer systems using ceramic monolith
substrates and fuel burners for regeneration. If these s.ame
traps were used with an electrical regeneration, the penalty
would likely be about the same, due to the energy required by
the alternator. However, if the ceramic fiber trap is used in
"the future, " the fuel economy penalty would be somewhat less.
The use of a catalyst to lower the ignition temperature of the
collected particulate would avoid the use of energy intensive
heating systems, since the traps would be self regenerating.
In this case, the fuel economy penalty would be lowered by an
amount that is basically equivalent to the energy used to
regenerate the other two trap systems. EPA estimates this is
equivalent to about a 0.5 percent fuel economy penalty.
Therefore, the use of a ceramic fiber trap might result in only
a 0.5-1.0 percent penalty rather than the 1.0-1.5 percent
penalty used in this analysis.
iii. Maintenance
The draft analysis identified two maintenance items for
trap-equipped HDDEs: the regeneration system and the exhaust
system. The regeneration system was assumed to require
maintenance after approximately five years of operation. At
that point, the engine temperature and trap temperature sensor
would need replacement. For the exhaust system, the customary
replacement of the standard steel exhaust pipe was expected to
-------
3-87
be eliminated, because this component would be displaced by a
stainless steel exhaust pipe as part of the trap-oxidizer
system. Sensor maintenance when discounted to the year of
vehicle purchase was estimated to range from $22 for LHDDEs
with one trap to $44 for HKDDEs with two traps. A net savings
was projected due to eliminating the need for exhaust pipe
replacement. This discounted savings was estimated at $39 for
LHDDEs, ranging up to $97 for HHDDEs.
Four general comments were received from metropolitan
transit authorities and the Department of Transportation
suggesting that maintenance costs would increase due to .the
combined NOx and particulate standards. Since the maintenance
savings that were projected for trap-based particulate
standards overwhelm the small cost associated with the NOx
standards, these comments would appear to be directed primarily
at the former standards. The New Jersey Bus Operations, Inc.,
specifically claimed that EPA's regulations will require the
use of electronic control units, resulting in substantial
expenditures for training, and the need for a more
sophisticated and expensive labor force. In another specific
comment, the Department of Energy estimated that the
particulate standards could save up to $402 for a HDDE. in
Classes IIB-VI and up to $519 for a HDDE in Classes VII-VIII
(undiscounted). Finally, a few manufacturers indicated in
their technical feasibility comments that a trap may
potentially require some type of maintenance during the
vehicle's lifetime due to such things as the accumulation of
ash or catalytic material.
In response to the concern expressed regarding the forced
use of electronic control units on urban bus engines.
electronics are projected to be widely used on all types of
HDDEs in the future regardless of emission control
requirements. Hence, costs associated with purported changes
in the labor force cannot be charged against the emission
standards. Concerning DOE's savings estimate and the general
comments that maintenance costs will rise, EPA will address
these comments by reevaluating the likely effect on maintenance
in. the context of the revised trap system design as described
in the section on variable costs (i.e., ceramic monolith/fuel
burner regeneration system).
The assumption in the draft analysis regarding sensor
replacement was not adversely commented upon, and is being
generally retained with a few revisions. The most significant
changes are the use of new component costs and a revision in
the assumed number of replacements for each HDDE size
category. The estimated retail cost of each engine temperature
sensor is $9 and each trap temperature is $20. [8] The
replacement of both sensors is estimated to take one hour at
-------
3-88
$28 per hour. To be consistent with the revised trap
configuration, all HDDEs are assumed to have a single trap ar.d,
therefore, only one trap temperature sensor is required per
affected vehicle. To account for the significantly different
lifetime mileages of the various HDDE classes, the number of
replacements over a vehicles life has been revised to one for
LHDDEs, two for MHDDEs, and three for HHDDEs. Using these
values and discounting the costs at 10 percent over the life of
the vehicle, the approximate cost is $35 for LHDDEs, $57 for
MHHDEs, and $71 for HHDDEs.
A review of EPA's previously estimated savings for exhaust
pipes has resulted in this category being eliminated to be
consistent with the revised ceramic monolith system design.
Again referring to the description contained in the variable
cost section, when the standard steel exhaust pipe is replaced
by its stainless steel counterpart (i.e., from the manifold to
^the bypass valve), the displaced standard pipe is assumed to be
'roughly equivalent to that required by the bypass system (i.e.,
from the bypass to the muffler). For the purposes of this
analysis, it is assumed that the replacement schedule of this
standard pipe will be roughly equivalent .regardless of
location. Hence, no incremental maintenance is estimated for
the exhaust system.
Regarding the possibility that traps could require some
type of maintenance, estimating any cost in this area' is
especially difficult due to the current limited information on
traps themselves. Nonetheless, to cover the potential costs of
such a contingency, EPA will assume that trap maintenance will
cost about $50 per event and that frequency of this maintenance
will parallel sensor maintenance. Therefore, when discounted
at 10 percent to the year of purchase, the cost is $31 for
LHDDEs, $50 for MHDDEs, and $62 for HHDDEs.
Total maintenance costs for the average ncn-bus HDDE with
traps is $102, while the cost for an urban bus is $107. On a
fleetwide basis, the cost per HDDE in the short-term is $72 and
declines in the long term to $62.
As in the fuel economy discussion, it is appropriate to
examine the potential maintenance costs associated with the two
other trap-oxidizer systems. The ceramic monolith/electrical
regeneration system would have the same maintenance
requirements and costs as the fuel, burner, since the same trap
substrate and sensors are used in both systems. If the ceramic
fiber trap-oxidizer is used in the future, the maintenance
requirements would be somewhat different. In addition to the
costs associated with sensor and potential trap maintenance,
catalytic material used in this system may need replacement
during'the vehicle's lifetime. Offsetting these costs would be
-------
3-89
a credit resulting from the use of the stainless steel exhaust
pipe, which eliminates the need for periodically replacing the
standard steel pipe. The discounted catalyst replacement
costs, as estimated using the basic methodology presented in
the variable cost section, are $14 for LHDDEs, $25 for MHDDEs,
and $86 for HHDDEs. The discounted exhaust pipe credits, as
estimated using the appropriate schedule for HDDV standard
steel pipe replacements described in the DPS are $41 for
LHDDEs, $51 for MHDDEs, and $81 for HHDDEs. Combining these
values with those previously estimated for sensor and trap
maintenance results in discounted ceramic trap maintenance
costs ranging from $39 to $138 depending on HDDE size. .For
LHDDEs this is significantly less than the costs estimated for
the ceramic monolith/fuel burner system, while for the largest
HDDEs it is about the same.
iv. Total User Costs
The total cost to the purchaser of an HDDE is composed of
the first price increase and the lifetime discounted costs for
a fuel economy and maintenance. The cost for each
trap-equipped non-bus HDDE is $577-604 for all LHDDEs-,
$901-1,030 for a MHDDE, and $1,499-1,852 for a HHDDE. For . the
average non-bus HDDE with • a trap the total cost is
$1,050-1,180. For an urban bus it is $1,712. Expressed as an
average over the entire fleet, the total user cost in the short
term is $723-843 and will decline to about $625-728 in the long
term. The fleetwide cost per vehicle is summarized in Table
3-36.
6. Total HDDE Manufacturer and User Costs for the 1991
NGx and "articulate Standards
The total cost to HDDE manufacturers of the 1991 standards
is the sum of the fixed and variable costs of the NOx and
particulate emission control. These costs are passed on to the
users of HDDVs as first cost increases, and are added to
operating costs for total user cost of the standards. These
values were developed above, and are presented in Tables 3-37
and 3-38.
The discounted manufacturer cost is about $476 million,
while the average increase in lifetime user cost for a 1991
model year HDDV is $803-1,171, tapering off to $700-977 for a
1993 model year HDDV.
7. 1994 Diesel Particulate Standard <0.10 g/BHP-hr
HDDEs)
In this section, the economic effects of the 1994
particulate standard for HDDEs are analyzed. The costs are
examined as an increment to those that would result from
-------
3-90
Table 3-36
Toe a I User Cost
for tr.e L99I Particulate Standards
(Discounted to Year of Vehicle Purchase)
Fleet Average Vehicle
Cost Category Short-Term Long-Term
First Cost $ 390 $ 336
Fuel Economy 261-381 227-330
Maintenance 72 62
Total $723-843 $625-728
-------
3-91
Table 3-37
Total HDDE Manufacturer Costs
1991 NOx and Particulate Standards
Und is counted
NOx
Particulate
Total
Grand Total
RD&T
$28. 7M
49. 5M
78.2-1
Hardware**
$42. 3M
371. 4M
413. 7M
$491. 9M
Discounted*
RD&T
$34. 8M
63. 4M
98. 2M
Hardware**
$38. 5M
339. 1M
377. 6M
$475. 9M
* Discounted at 10 percent to 1991
** Model year 1991-93 HDDVs.
-------
3-92 -
Table 3-38
Total HDDE User Cost*
1991 NOx and Particualte Standards
NOx _ .
Particulate
Total
Grand Total
First
Cost
$68
$390
458
Short Term
Long Term
Fuel Maint-
Economy
$0-348
261-381
261-729
791-1,259
enance
$0
72
72
First
Cost
$68
336
404
Fuel
Economy
$0-174
227-330
227-504
693-970
Maint-
enance
$0
62
62 -
Incremental cost over cost of 1988 standards.
-------
3-93
continuing the 1991 standard into the 1994 and later model
years. As discussed in the Technical Feasibility Chapter, tne
Agency expects that the 1991 0.25 g/BHP-hr standard would
require about 60 percent of the non-bus HDDEs to be trap
equipped in the long term (i.e., about 1994). Similarily, the
1991 0.10 g/BHP-hr standard would require all urban bus engines
to be trap equipped. With the more stringent 1994 0.10
g/3HP-hr, about 90 percent of the non-bus HDDEs will need
traps, while the buses will all remain trap equipped.
Therefore, the incremental effect and resulting cost of the
1994 standard is dependent on the use of an additional 30
percent traps by non-bus HDDEs.
The basic inputs for this analysis are taken from the 1991
particulate standards section where the costs of various
trap-oxidizer systems were examined. Specifically, that
analysis reviewed the costs associated with a ceramic monolith
trap using a fuel burner regeneration system, a ceramic
monolith trap using an electrical regeneration system, and a
ceramic fiber trap using a catalyst regeneration system. The
details of that comprehensive evaluation will not be repeated
here. It is important to note, however, that only the ceramic
monolith/fuel burner system was used to estimate the econgmic
effects of the 1991 standards. This approach was taken, in
spite of the potentially lower cost of the ceramic
fiber/catalyst system, because the ceramic monolith/fuel burner
trap is presently the most well defined and may be the first
commercially available trap-oxidizer.
The economic effects of the 1994 standard will also be
assessed using the costs associated with the ceramic
monolith/fuel burner system. Due to the long leadtime
associated with the 1994 requirement, however, it is possible
that a lower cost trap oxidizer such as the ceramic
fiber/catalyst system may be widely used by the effective date
of the standard. If this were to occur, the cost of the 1994
standard would be somewhat less than that presented in the
subsequent sections. - -
a. Cost to the Manufacturers
i. Fixed Costs
The total fixed cost of the 1994 standard obviously will
be significantly less than that associated with the 1991
standards. Only 30 percent of the HDDEs will incur development
and certification testing costs, compared to about 70 percent
in 1991. Also, when reviewed on a per vehicle basis, it is
reasonable to expect that engineering experience gained
throughout the early 1990's will make the application of traps
to new families in 1994 less difficult than it was in 1991.
-------
3-94
Therefore, the fixed cost that was recovered in the sales price
of a trap-equipped HDDE under .the 19S1 standards (i.e., about
$87) would seem to represent an upper limit for the fixed cost
associated with the 1994 standards.
Using this conservative assumption, the total fixed cost
of the 1994 standards can be estimated by multiplying $87 per
trap-equipped vehicle by the number of such vehicles over which
fixed costs are recovered. As in the previous HDDE analyses,
the fixed costs of a standard are assumed to be recovered over
three years of production immediately following the effective
date of the standard. The estimated cost of capital is 10
percent per annum. Based on the projected HDDE sales in Table
3-11, the number of trap-equipped HDDEs used in the fixed cost
calculation is about 334,000 (i.e., 30 percent of the 1994
through 1996 HDDE "discounted sales," excluding buses). This
results in total estimated fixed costs of $29.1 mi 11 ion, """when
expressed as a lump sum investment in 1994. The expenditures
-of these over time can be expected to occur as shown in Table
3-39.
ii. Variable Costs
The variable cost of a specific trap-oxidizer system is a
function of vehicle size. The costs of a ceramic monolith trap
with a fuel burner regeneration system was previously estimated
at $370 for LHDDEs, $448 for MHDDEs, and $574 for HHDDEs.
Total HDDE sales, excluding urban buses, are composed of 36
percent LHDDEs, 27 percent MHDDEs, and 37 percent HHDDEs.
Using these values to sales weight the trap-oxidizer cost for
each size category results in an average variable cost of $4-67
per trap-equipped HDDE. Expressed as an average over the
entire fleet, the variable cost is $137 per HDDE.
iii. Total Manufacturers Cost of the 1994 Particulate
Standards
The total discounted and undiscounted costs to
manufacturers are shown in Table 3-40. The fixed costs are
taken directly from Table 3-39. The variable costs are the
products of the hardware cost per fleet-average vehicle and
total HDDE sales in each specific year (Table 3-11). The total
undiscounted cost is $193.5 million, while the discounted cost
is $217.9 million.
b. Cost to Users for HDDEs Complying with the 1994
Particulate Standards
i. First Cost
The amount that a manufacturer must increase the price of
an HDDE to recover its expenses depends on the timing of the
costs, the cost of capital, and the number of vehicles over
-------
Taole 3-39
Years
1990
1991
1992
1993
Total
Torrai Fixed Cose of the
1994 HDDE Parciculate Standard
Discounted
Fixed Costsfl]
$ 5.0 M
13.0 M
7.9 M
3.2 M
Undiscounted
Fixed Costs
$3.4 M
9.8 M
6.5 M
2.9 M
$29.1 M
$22.6 M
"[Tl Discounted to the effective date of the standard,
1994.
i .e.,
-------
Tacle 3-40
HDDE .".anu£-3Cturers ' Cose
for the !?">•> Particulate Standard
Undiscounted Discounted
Year Fixed Cost Variable Cost Total Total
1990 $3.4 M — $ 3.4 M $ 5.0 M
1991 9.8 M — 9.3 M 13.0 M
1992 6.5 M — 6.5 M 7.9 M
1993 2.9M — 2.9M 3.2M
1994 — $56.0 M 56.0 M 56.0 M
1995 — 57.0 M 57.0"M 51.8 M
1996 — 57.9 M 57.9 M 47.8 M
Total • $193.5 M .. $184.7 M
-------
3-97
which the fixed costs will be recovered. As discussed in
deriving the fixed costs, rranufacturers are expected to recover
fixed costs over the first three years of production. The cost
of capital was also identified as 10 percent per annum. Hence,
the first cost increase for a vehicle is the sum of a portion
of the discounted fixed cost and the hardware cost, as
described earlier. Using this methodology, the purchase price
increment for a trap-equipped HDDE is estimated at $457 for
LHDDEs, $535 for MHDDEs, and $561 for HHDDEs. This averages
$553 per trap-equipped MODE. Expressed on a fleetwide basis,
the purchase price increase is about $163.
ii. Fuel Economy
Traps may adversely affect fuel economy due to a potential
increase in exhaust backpressure and because of the energy
required to initiate regeneration. The penalty associated with
the use of this technology was estimated in the 1991
particulate standards discussion as- 1.0-1.5 percent per
trap-equipped vehicle. When discounted to the year of vehicle
purchase, this is equal to approximately $54-81 for a LHDDE,
$259-388 for a MHDDE, and $705-1,058 for a HHDDE. This amounts
to $350-525 for the average trap-equipped HDDE. For .the
fleet-average HDDE, the discount lifetime fuel penalty is about
$103-154.
iii. Maintenance
The potential maintenance costs associated with trap
oxidizers fall primarily into two categories: sensor
replacement and trap maintenance. The discounted lifetime
costs associated with these items were estimated in the section
on the 1991 particulate standards as being $66, $107, and $133
for a LHDDE, MHDDE, and HHDDE, respectively. For the average
trap-equipped HDDE this is $102. Expressed on a fleetwide
basis, the discounted lifetime maintenance increment is
estimated $30.
iv. Total User Costs
The total cost to the purchaser of an HDDE is composed of
the first price increase, and the discounted lifetime costs for
fuel economy and maintenance. The total user costs for
trap-equipped HDDEs complying with the 1994 standard • are
$120-147, $366-495, and $838-1,119 for HHDDEs, LHDDEs, and
MHDDEs, respectively. For the average HDDE with a trap, the
total cost if $1,005-1,181. Expressed as an average over the
entire fleet, the total user cost is $296-347. The various
fleetwide costs per vehicle are summarized in Table 3-41.
-------
Taoie 3-41
Total '-'sec-Cose for ens
1994 Particulate Standard
(Discounted co Year of Vehicle Purchase)
Cost Category Fleet-Average Vehicle
First Cost $163
Fuel Economy 103-154
Maintenance 30
TOTAL $296-347
-------
3-99
8. Aggregate Costs to the Nation of the HDE NOx and
Particulate Standards
The aggregate costs to the nation of the HDE NOx and
particulate standards include the total manufacturer costs of
RD&T and hardware, and user costs of fuel economy and
maintenance which will be incurred due to the more strict
emission control requirements of the standards. These costs
were developed above, and are shown in Tables 3-42 and 3-43
according to the year of expenditure. All costs before the
year of the standard are for RD&T, including certification, and
costs after the year of the standard are for hardware and
additional operating costs for the venicles equipped with HDEs
projected to be sold in those years.
The aggregate costs presented in Tables 3-42 and 3-43 for
each model year group are incremental in nature. The aggregate
incremental costs for the 1991 model year group represent only
the added costs beyond those incurred in the 1988 model year
group. The same is true in considering the 1994 model year
group aggregate costs, with the exception that the increment is
calculated relative to 1991.
All costs are shown undiscounted in Table 3-42 and
discounted at 10 percent to the year of the standard in Table
3-43 and are developed in the preceding sections. As shown,
the aggregate costs to the nation of the HDE NOx and
particulate standards are approximately $118-600 million for
the 1988 standards, $833-1,241 million for the 1991 standard,
and $336-394 million for the 1994 particulate standards,
discounted to each of those years.
-------
3-100
D. Socioeconomic Impacts
The socioeconomic impact section in the Draft RIA
discussed the effects on manufacturer sales and cash flow, the
regional effects of employment, and the national effects on
vehicle purchasers, energy usage, balance of trade, and
inflation. These effects will not change significantly as a
result of the reanalysis of costs, since cost estimates
decreased or rose only slightly from the original estimates.
However, some comments were received from citizens,
environmental groups, the American Trucking Association (ATA),
and public transit system operators concerning the
socioeconomic ' impact of costs on individuals and
organizations. The questions raised by these comments are
reviewed in the following paragraphs.
Comments received from citizens and environmental groups
argued that the cost of these regulations are rightly passed on
to the consumers who also receive the benefits of improved
-------
3-101
Table 3-42
Undisccunted Aggregate Increrer.tal
Costs 3t the HDE Standards
(millions of dollars)
Model Year 1988
Calendar
Year
1986
1980
1988
1989
1990
HDDE NOx HDDE Particulate
$18.0
12.7
10.8-246.0
11.2-195.5
11.7-107.5
$16.0
7.0
5.9
6.2
6.4
Model Year 1991
Calendar
Year
1987
1988
1989
1990
1991
1992
1993
Model Yea
Calendar
Year
1990
1991
1992
1993
1994
1995
1996
HDDE NOx HDDE Particulate
_
$7.0
15.0
6.7
13.7-146.6
14.2-117.8
14.4-84.5
r 1994
HDE Particulate
$3.4
9.8
6.5
2.9
110.4-131.3
112.3-133.6
114.2-135.7
$8.0
20.0
14.0
7.5
253.9-299.8
253.0-298.6
239.8-282.4
HDGE NOx
$1.6
2.1
1.3
1.3
1.3
HDGE NOx
$3.5
2.1
2.8
2.8
2.8
HDE Total
$35.6.
21.8
18.0-253.2
18.7-203.0
19.4-115.2
HDE Total
$8.6
27.0
32.5
16.3
270.4-449.2
270.0-419.2
259.0-369.7
-------
3-102
Model Year 1988
Table 3-43
Discounted* Aggregate Encrerrental
Costs oc the HDE Standards
(millions of dollars)
HDGE NOx
Calendar
Year
1986
1987
1988
1989
1990
Total
Model Year
Calendar
Year
1987
1988
1989
1990
1991
1992
1993
Total
Model Year
Calendar
HDDE NOx HDDE Particulate
$21.8
- 14.0
10.8-246.0
10.2-177.7
9.7- 88.8
$66.5-548.3
1991
$19.4
7.7
5.9
5.6
5.3
$43.9
HDDE NOx HDDE Particulate
,
$9.3
18.2
7.4
13.7-146.6
12.9-107.1
11.9- 69.8
$73.4-358.4
1994
$11.7
26.6
16.9
8.2
253.9-299.8
230.0-271.4
198.2-233.4
'$745.5-868.0
Year HDE Particulate
1990
1991
1992
1993
1994
1995
1996
Total
$5.0
13.0
7.9
3.2
110.4-131.3
102.0-121.3
94.3-112.1
$335.8-393.8
•
•
HDE Total
$43.1
24.0
18.0-253.2
17.0-184.5
16.1- 95.2
$113.2-600.0
HDGE NOx
HDE Total
$11.7
35.9
39.4
17.9'
270.4-449.2
245.4-381.0
212.4-3Q5.5
$833.1-1240.6
10 percent to r.r\e year of the standard.
-------
3-103
environment and public health, and thus "will think that it is
worth the cost." The Agency agrees with these corrjnents. EPA
expects the manufacturers to recoup their losses through first
price increases in LDTs and HDEs.
On the opposite side of the argument, ATA believed that
the costs may be too high, stating that, "at issue is not
whether the average motor carrier will be adversely impact but
rather, in the case of fuel penalties for example, the
magnitude of this effect at the upper end of the" range in
potential penalties on the highest mileage group of single
truck or small fleet owner/operators." In response, the Agency
believes that the ATA has posed an unrealistic scenario. There
is no reason to expect the maximum operating cost impact to
fall on small high-mileage operators, since these operators
"will certainly search the market for the vehicles with minimal
fuel economy impact if operating cost is of great concern. A
comment by the National Resources Defense Council is relevant
here, which states that, "even the more expensive standards
still add only a small fraction to the initial cost and
lifetime operating cost of the vehicles in question." Costs
should be able to be easily borne by the trucking industry with
small increases in the prices of consumer goods; since these
costs will be carried by all segments of the industry, no one
group should receive an unfair advantage or disadvantage due to
the standards.
Several comments were received pertaining specifically to
. the socioeconomic impacts of the proposed NOx and particulate
standards on urban transit Anises. The Urban Mass
Transportation Administration (UMTA) and local transit and
transportation authorities from New Jersey, Washington,
Chicago, Cleveland, Washington, D.C., Albany, and San Antonio
all stressed the economic burden that would be placed upon
urban transportation locally and nationally. There was general
agreement among these agencies that EPA underestimated the true
costs and economic burdens associated with the proposed
standards. New Jersey Transit and VIA Metropolitan Transit in
San Antonio indicated that the increased costs would be
translated into higher* fares, lower riderships, more personal
vehicle use, and an increase in emissions as a net result of
the proposed standards. Finally, the Chicago Transit Authority
(CTA) expressed concern that engine selection for transit buses
would be reduced as manufacturers leave the market due to the
increased costs of control.
EPA has estimated the first price increase associated with
a 0.10 g/BHP-hr particulate and 5.0 g/EHP-hr standard at a
value of $644. The total fuel economy penalty resulting from
these controls is estimated to be 2 percent, or $1427, in the
long run, and slightly higher in the short run. There is also
-------
3-104
a rsaintenance cost cf $107 per bus associated with the
particulate standard. Wich the current average price of a
diesel transit bus being $135,000-145,000, the first price
increase estimated represents at most a 0.5 percent increase in
the first price of a diesel transit bus. The operating and
maintenance cost associated with an urban transit bus will rise
at most slightly over 2 percent. This assumes that fuel is the
only operating cost involved; other considerations would reduce
this figure. Thus, the "economic burden" associated with the
NOx and particulate standards does not appear to EPA to be
severe. Based on this, EPA does not believe that there will be
any significant fare increases and associated ridership losses
attributable to the standards.
The market for diesel engines used in transit buses is
small as CTA has indicated. Currently only one domestic
manufacturer, GM, makes engines for large urban transit buses,
and only one or two of their five such engines are made
expressly for that purpose. Also, Caterpillar makes an engine
used in smaller transit buses in some urban areas. EPA feels
that it is highly unlikely that either manufacturer would
relinquish its share of the market under such circumstances.
-------
3-105
References:
1. "Cost Estimations for Emission Control Related
Components/Systems and Cost Methodology Description," Rath and
Strong Inc., EPA-460/3-78-002, March 1978.
2. "Diesel Particulate Study," U.S. EPA, OAR, QMS,
ECTD, SDSB.
3. "Heavy-Duty Vehicle Emission Conversion Factors
1962-1997," Mahlon C. Smith IV, EPA-AA-SDSB-84-1, August 1984.
4. "Standards for Emission of Particulate Matter From
Diesel-Powered Light-Duty Vehicles and Light-Duty Trucks; Final
Rule" (49 FR 14496, March 5, 1980).
5. "Standards for Emission of Particulate Matter From
Diesel-Powered Light-Duty Vehicles and Light-Duty Trucks and
Technical Amendment to Emission Regulations for Light-Duty
Vehicles, Light-Duty Trucks, and Heavy-Duty Engines; Final
Rule" (49 FR 3010, January 24, 1984).
6. Oral Communication with Terry Ullman, Southwest
Research Institute, San Antonio, Texas, January 28, 1985.
7. "Cost of Selected Trap-Oxidizer System Components
for Heavy-Duty Vehicles," Jack Faucett Associates and Mueller
Associates, September 28, 1984.
8. "Costs of Selected Heavy-Duty Diesel Engine Emission
Control Components," Jack Faucett Associates and Mueller
Associates, February 8, 1985=
9. "1983 and Later Model Year Heavy-Duty Engines,
Proposed Gaseous Emission Regulations: Summary and Analysis of
Comments to the NPRM," EPA, OMS, ECTD, December 1979.
10. "Cost Estimations for Emission Control Related
Components/Systems and Cost Methodology Description, Heavy-Duty
Trucks," Rath & Strong,- Inc., EPA-460/3-80-001, February 1980.
11. "Light-Duty Diesel Organic Particulate Control
Technology Investigation," Southwest Research Institute,
EPA-460/3-82-011, August 1983.
12. "Control of Air Pollution From New Motor Vehicles
and New Motor Vehicle Engines; Federal Certification Test
Results For 1984 Model Year," U.S. EPA, OMS, CD.
-------
3-106
13.
Diesel Bus
Thomas M. Baines,
Preliminary Particulate
Engine," Terry L. -Ullnan, Charles
SAE No. 840079, February 27,
Trap Tests on
1984
14. Memo from Charles L
Request Regarding a Manganese
Trap Regeneration System, U.S.
. Gray, Jr. to Robert
Fuel Additive-Based
EPA, QMS.
a 2-Stroke
Hare, and
Maxwell, VW
Particulate
-------
CHAPTER 4
MGx AND FARTICULATE ENVIRONMENTAL IMPACT
This chapter will examine the environmental effects which
can b.e expected to result from the implementation of the
revised NOx standards for light-duty trucks and heavy-duty
engines and new diesel particulate standards for heavy-duty
diesel engines. The material presented here begins with an
overview of Chapters 4 and 5 in the Draft Regulatory Impact
Analysis, followed by a summary and analysis of the comments
made on the information contained in these chapters, and,
finally, a presentation of revised projections of the
environmental and air quality impacts of the NOx and diesel
particulate emissions.
I. Overview of NPRM Analyses
A. Oxides of Nitrogen (NOx)
The Draft analysis opened with a brief review of the
health effects associated with NOx emissions. The primary
concerns reviewed were the human respiratory effects which
formed the basis for the level of the primary ambient N0'2
standard. At the present time, this standard level is an
annual arithmetic mean of 0.053 ppm.
Following this review, the effect of the proposed NOx
standards on ambient air quality was estimated by comparing
future year NOx emissions inventories and ambient NO2 levels
under three scenarios: 1) no future control, 2) the proposed
standards, and 3) the eventual standards as mandated in the
Clean Air Act. These analyses focused on those urban areas
that are wichin range of exceeding the NAAQS by the end of the
century. In addition, estimates of lifetime emission
reductions per vehicle were made, primarily for use in the cost
effectiveness analysis.
The air quality analyses for NOx were performed using a
three-step approach. The first step involved the use of
MOBILE2.5 to estimate emission factors by calendar year and
vehicle class under the three scenarios. MOBILE2.5 determines
emission factors in grams per mile (g/mi) for motor vehicles,
based upon vehicle class, engine type, model year, and age of
the vehicle. For heavy-duty engines, additional factors are
used to convert brake-specific emission factors to g/mi
emission factors. In order to obtain a specific calendar year
emission factor for the individual vehicle classes,
dieselization rates by model year, registration distributions
by age, and mileage accumulation rates by age are combined with
the emission factor by model year and age. The model year
-------
4-2
emission factors reflect inproverrents in control efficiency
over time. The calendar year emission factors are then
utilized by the EPA Rollback Model in step 3, described below.
In the second step, base year inventories of NOx emissions
for the urban areas of interest were obtained from the National
Emissions Data System (NEDS).[1] NEDS provides county-specific
estimates of emissions by source category for each county in
the United States. Total vehicle miles travelled (VMT) by
county, VMT breakdown by vehicle class by county, and vehicle
emission factors are the key parameters in determining the
mobile source inventory. The 1981 NEDS inventory contained, in
the draft analysis was derived using emission factors from
MOBILE2.
These were combined, along with current NOZ levels and
projected growth in source activities and control efficiencies,
to yield future year emissions and NO* levels. This final
.step was performed using the EPA Rollback Model which begins
with base year inventories of NOx emissions' and base year
ambient levels of NOZ concentration (design values).
Utilizing the emission factors from the MOBILE program, along
with projections of total VMT by vehicle type, and similar
numbers for stationary sources, the model can then project
future year inventories of NOx emissions and corresponding
ambient levels of NOZ. The emissions from the various
sources are discounted to reflect their impact upon air qual.ity
in the immediate local area. Increases in ambient NO? levels
are assumed to move linearly with increases in discounted NOx
Remissions. ,
Estimates of lifetime reductions in NOx emissions per
vehicle were calculated in a straightforward manner.
Differences in the emission factors by mileage for the various
control scenarios and estimates of mileage accumulation over
time for the appropriate vehicle classes (obtained from
MOBILE2.5) were combined and summed over the vehicle's life.
A more complete description of the modelling procedures
can be found in the Draft RIA, and in the following documents:
"User's Guide to MOBILE2",[2] "Compilation of Air Pollution
Emission Factors: Highway Mobile Sources",[3] and "Rollback
Modelling: Basic and Modified".[4]
B. Particulate Matter
The Particulate Environmental Impact Chapter in the Draft
RIA opened with a discussion of the relationship of diesel
particulate matter to total suspended particulate and the NAAQS
for particulate matter. The widespread non-attainment of the
NAAQS in 1995, under either the current TSP standard or the
proposed PM10 standard, was emphasized.
-------
Following this discussion, the lifetime reductions in
particulace emissions per vehicle were then derived, a'-ain for
use in the cost-effectiveness analysis. These reductions were
estims-ted using the same basic methodology as that described
above for the NOx analysis.
Next, nationwide and nationwide-urban emissions of diesel
particulate were presented. These projections were made using
the same basic methodology as for NOx, but with slight
modifications. For instance, due to the widespread violation
of the particulate NAAQS, it is not reasonable to model each
urban area individually. Thus, all U.S. urban areas were
analyzed together. Also, the MOBILE model itself is not
equipped to determine emission factors for diesel particulate
matter, so it could not be used in the diesel particulate
analysis. However, the concepts of MOBILE and all applicable
parameters contained in MOBILE2.5 (described in detail in the
Diesel Particulate Study, or DPS[5]) were used to estimate
calendar year emissions.
Since the diesel particulate analysis is done on a
nationwide, and not on an individualized urban area basis, NEDS
is not used as the source for the base year inventories.
Instead, emission factors were combined with estimates of
nationwide urban VMT by vehicle class to develop base year
inventories of diesel particulate emissions.
The estimates of nationwide emissions were then followed
by projections of ambient diesel particulate levels. Due to
- the difficulties in distinguishing diesel particulate from
others in atmospheric measurements, some measurable surrogate
in ths ambient air that is directly relatable to vehicular
emissions must be used to estimate current ambient diesel
particulate levels. The two surrogates that have historically
been used are lead and CO. Three types of ambient impact were
addressed: 1) levels expected to occur at air quality
monitors, 2) average exposure levels of urban dwellers, and 3)
ambient levels in selected high-exposure situations.
* >.
In estimating urban monitor-type levels, conceptually,
historic ambient lead levels are first converted to historic
ambient diesel particulate levels. This is done by assuming
that the ratio of ambient concentrations of the two pollutants
is equal to the ratio of their emissions, taking into account
that a certain fraction of leaded particulate emitted falls out
of the atmosphere very quickly and does not affect ambient air
quality. Future ambient diesel particulate levels are then
projected from historic levels using the general rollback
approach. Projections were made for a broad spectrum of city
sizes and meteorological conditions.
-------
4-4
Annual average urban exposures, which include a vanecy of
individual activity pattern elfects, were based on a model
developed by EPA to estimate exposures under various levels of
the CO NAAQS. The model was based on measured exposures in
specific types of situations in four U.S. cities, and involved
placing the population into various cohorts which spend various
amounts of time in each exposure situation. The CO levels
projected by the model were converted to diesel particulate
analogously to the conversion described above for the
lead-surrogate model.
»The high-exposure, or microscale, situations were analyzed
using models developed for EPA for the projection of any
completely dispersed, non-reactive pollutant. Thus, they are
also based on the surrogate and rollback concepts. Four
"situations were modelled: roadway tunnels, street canyons, on
an expressway, and nearby an expressway.
Following these three estimates "of microscale
concentrations of diesel particulate, the particular need to
control diesel particulate at high altitude was discussed.
While the lack of particulate emission data at high altitude
prevented any more precise estimate of environmental impact
than that presented in the nationwide analysis described above,
- Denver's air quality situation was discussed briefly and the
need for high-altitude control was established.
Following these emission and air quality analyses, the
Draft RIA attempted to put these projections in perspective by
—examining four classes of health and welfare effects associated
'"•'with diesel particulate: 1) non-cancer health effects, 2)
carcinogenic health effects, 3) visibility, and 4) soiling.
The analysis of non-cancer health effects associated with
diesel particulate focused on identifying the potency of diesel
particulate relative to that of general suspended inhalable
particulate (i.e., PMi0). Using this relative potency, the
ambient diesel particulate levels identified earlier were
compared to the current PM10 levels of urban areas and the
proposed PM10 standards.
With respect to carcinogenic effects, an estimate of the
lifetime risk of contracting lung cancer from exposure to
diesel particulate was made using estimates for the potency of
diesel particulate and the earlier estimates of average urban
exposure. Due to the limited epidemiological data available,
the estimate of the carcinogenic potency of diesel particulate
was made using a comparative potency method developed by
EPA.[6] In this methodology, the relative potency of diesel
particulate to known human carcinogens is determined from the
relative potencies of the compounds in non-human laboratory
bioassays and then applied to known human cancer risks of the
human carcinogens.
-------
4-5
As the size ar.d cherr.icsi corposition of diesel particulate
rr.akes it very effective in boch scattering and absorcing light,
a was developed to quantify the reduction in visibility caused
by a.Tt.ient diesel particulate levels in a large number of urban
areas.t5] The model used the projections of ambient diesel
particulate levels described earlier, Beers' law, a measured
coefficient of extinction for diesel particulate, and the
assumption that diesel particulate levels were constant inside
the city radius and zero outside the radius to determine the
visibility reduction.
The effects of soiling due to diesel particulate are
described briefly in the Draft RIA. Little physical data were
found describing the rate of particulate soiling or the soiling
of diesel particulate relative to that of other types.
However, due to its black color and oily nature, diesel
particulate may have a disproportionate effect on soiling
compared to the effect of other types of particulate matter.
The only quantitative estimates of soiling were economic in
nature and made in Chapter 8 of the Draft RIA (Cost-Benefit
Analysis).
A more complete description of the methodologies described
above can be found in the Draft RIA and the DPS.[5]
II. Summary and Analysis of Comments on NPRM Environmental
Impact and Air Quality Projections
Numerous comments were received from vehicle and engine
manufacturers, public transit organizations, environmental
groups and private citizens, dealing largely with various
specific inputs used to project future emissions and air
quality in the NPRM analyses. Several of the issues addressed
are common to both the NOx and diesel particulate analyses, and
will be dealt with in the first part of this section. This
discussion of common parameters will be followed by two sets of
discussions dealing with factors specific only to the NOx and
particulate projections, respectively.
A. Factors Common to Both Analyses
1. Baseline VMT Breakdown
A critical parameter in estimating both NOx and
particulate emissions is the breakdown of VMT by vehicle class
in the area being examined. These VMT breakdowns were under
study by EPA just prior to the issuance of the NPRM. At that
time, it was discovered that the VMT breakdowns used in the NOx
projections, which were taken from the National Emissions Data
System (NEDS)[1] for selected SMSAs, were quite different from
the "Nationwide Urban" VMT breakdown used in the particulate
analysis, which was developed primarily from the Energy and
-------
4-6
Environmental Analysis, Inc. (EEA) fuel consumption model.[7]
At the hearing following the NPRM, a technical report entitled
"Motor Vehicle NOx Inventories"[8] was issued showing that the
"Nationwide Urban" approach allocated a significantly lower
percentage of total urban VMT to heavy-duty diesel vehicles
(HDDVs) than did the NEDS methodology. Investigation into the
NEDS method of county-by-county allocation of statewide VMT
revealed some likely inaccuracies, especially with respect to
an overestimation of HDDV VMT in urban areas. The suspected
overestimation by NEDS was confirmed by estimates gathered from
local transportation and planning authorities, which on average
indicated a HDDV fraction of VMT very close to that estimated
using the "Nationwide Urban" approach.[8]
Comments received on the base-year VMT breakdown used in
the NOx projections and the above-mentioned technical report
indicated support for the use of the local transportation
agency data from each of the cities being modelled for NOx
emissions. However, as the technical report explained, local
data were available for only seven of the eleven cities in the
NOx analysis. The use of updated 1981-83 average NO2 design
values (discussed below) resulted in the introduction of three
new cities into the NOx analysis for which no local estimates
have been obtained and the removal of two cities for which
estimates were available. Thus, local data are now available
for only a minority of the cities being modelled. To further
complicate matters, subsequent analysis uncovered inaccuracies
similar to those found with the NEDS approach in two of the
seven available local estimates.[9]
*.-»"-
Therefore, both the NOx and diesel particulate projections
presented in this final rulemaking are based on VMT breakdowns
by vehicle -class developed using the "Nationwide Urban"
approach, which are very similar to the average of the local
data which are available and contain no known errors. This
method provides the flexibility needed to accommodate ongoing
changes in the cities being analyzed, yet addresses the largely
non-urban nature of HDDV travel (an improvement over NEDS).
Because the "Nationwide Urban" approach has been updated to be
consistent with MOBILES (the model used is called the MOBILES
Fuel Consumption Model)*, the breakdown of VMT by class is
The MOBILES Fuel Consumption Model (M3-FCM) is a recently
developed model, similar in principal to EEA's model,
which estimates nationwide and urban VMT and fuel usage by
vehicle class and fuel type. EPA's model is based
primarily on MOBILES fleet characterization data (from
NPTS and TIUS) and uses historic trends in vehicle
registrations (from R.L. Polk) to project future VMT
(mileage/vehicle assumed to be constant over time). Urban
VMT fractions and gas/diesel sales splits used in the
model are those presented in Tables A-2 through A-5 of the
Appendix.
-------
4-7
slightly different than that shown in "Motor Vehicle NOx
Inventories"; however, the basic rrethodclogy and the urban
fractions of VMT for each vehicle class are essentially tne
same, while only the nationwide VMT breakdown by vehicle class
differs. In particular, the resulting HDDV fraction of urban
VMT is essentially the same as that with the nationwide
approach presented at the hearing and the average of the
available local data. (Annual VMT by vehicle class, as
estimated by the 'MOBILES Fuel Consumption Model and used in the
final analyses, is presented in Table A-l of the Appendix. The
urban fractions of VMT used are shown in Tables A-2 and A-3 for
heavy-duty diesel and gas vehicles, respectively.[10] Urban
fractions of LDV and LDT travel are assumed to remain constant
over time at 0.597 and 0.514, respectively, based on 1983 FHwA
-data.[11])
Final estimates of 1982 urban VMT breakdown by class, used
in both the final NOx and diesel particulate analyses, are
presented below: -- - - — .. ._
Vehicle Class % of Total 1982 Urban VMT*
Light-duty Vehicle (LDV) 72.8
- Gasoline (71.2)
- Diesel (1.6)
Light-duty Truck (LDT) 20.5
- Gasoline (20.1)
- Diesel (0.4)
Heavy-duty Gas Vehicle (HDGV) 4.4
Heavy-duty Diesel Vehicle (HDDV) 2.3
Total 100.0
These percentages, applied to 1962 VMT totals and then
multiplied by 1982 NOx and diesel particulate emission factors,
were used to develop base-year pollutant inventories for the
emissions projections presented later in this chapter.
2. VMT Growth Rates
A modelling parameter that received a substantial amount
of comment was the set of VMT growth rates that were applied to
base-year VMT for each vehicle class to project future VMT.
Specific recommendations concerning the appropriate levels of
VMT growth were submitted by General Motors (with support from
other manufacturers) and DOE (quoting EEA-based figures).
Comments were also received from the American Trucking
Because of the use of updated NO2 design values (to be
addressed later in this chapter), an update from 1981 to
1932 base-year VMT was necessary.
-------
4-8
Association (ATA), stating that future VflT by HDDVs will be
reduced due to the replacement of some conventional
truck-trailer combinations with twin trailers (i.e., cne
tractor pulling two trailers). Although ATA carr.e to no final
conclusion on an appropriate HDDV growth rate, Argonne National
Laboratory was cited as a reliable independent source. As the
Argonne VMT model (TEEMS) is being considered for use in the
Federal government's National Acid Precipitation Assessment
Project, and recent output of the model was available,
Argonne's independent projections of future VMT growth are
included for purposes of comparison in this analysis.[12]
In general, GM's estimates for each of the vehicle classes
are lower than the growth rates used in the NPRM projections
and lower than those recommended by both DOE and Argonne.
Table 4-1 summarizes the VMT growth rates suggested by the
commenters (along with Argonne), compared to the rates used in
.the NPRM analyses and those chosen for the FRM projections.
The final (FRM) growth rates shown in Table 4-1 are based
on urban VMT projections made using the MOBILES Fuel
Consumption Model (M3 FCM), calculated from the VMT figures
shown in Table A-l. This is the same model used to develop-the
base-year urban VMT breakdown by vehicle class. The growth
rates are nationwide averages for urban areas across the U.S.;*
city-specific growth rates were not determined for the same
reasons given earlier in the base-year inventory discussion —
absence of local projections from some cities and need to
accommodate changes in the specific cities being modelled.
As Table 4-1 shows, the FRM (M3 FCM) growth rates for the
LDV and LDT classes are in basic agreement with Argonne's
independent projections, estimating LDT growth at a slightly
higher level than LDV growth. The LDT growth rate is
significantly lower than that used in the NPRM analysis, which
was based on EEA's Eighth Quarterly Report.[13] GM's
projections also show equal rates for LDVs and LDTs. However,
their light-duty growth rates are significantly lower than the
other estimates, most likely due to GM's assumption that both
LDV and LDT VMT growths are primarily a function of growth in
U.S. population. Although GM does state that there were
adjustments made to account for trends in per-capita vehicle
ownership and in miles driven by individual vehicles,[14] their
approach still appears to underestimate future light-duty VMT
growth in comparison with independent projections from both
Argonne and EEA, based on more sophisticated econometric models.
In addition to urban VMT growth rates, nationwide growth
rates were also calculated from the M3 FCM for use in the
NOx analysis; both the urban and nationwide growth
estimates are shown in Table A-7 of the Appendix.
-------
4-9
Tacie 4-1
Annual Compound Uroan VMT Growth Rates
(Percent per Year)
Vehicle
Class
LDV
LOT
HDGV
HDDV
HDV (overall)
NPRM
+ 1.7
+4.7
-0.3
+ 6.4
— _
EPA
Interim
Ana Ivsis
+2.0
+4 .0
+ 2.1
+ 6.7
— —
GM
+ 1.2
+ 1.2
-2.6
+ 3.6
+ 1.1
DOE Araonne FR.M
-- . +1.9 +1.9
— • +2.3 +2.1
+ 0.6
+ 6.9 — +4.2
+2.0 +2.0
Note:
NPRM — Based on EEA's Eighth Quarterly Fuel
Consumption Model Report with assumptions;
1980-1995.
EPA Interim Analysis— Based on EEA 10th Quarterly Report,
with urban assumptions from TIUS and FHwA;
1981-1995. _
GM —
DOE --
Argonne--
FRM--
Based on
1978-2000.
1980
OBERS with assumptions;
Based on EEA data and projections; 1980-1995.
Based on ANL-83N forecast, TEEMS; NAPAP likely
to be similar; nationwide estimates; 1980-2000.
Based on MOBILES Fuel Consumption Model;
1982-2000.
-------
4-10
With respect to overall heavy-duty growth, GM's estimate
is based on 1980 Department of.Commerce (DOC) OBERS projections
for future growth in employment within the construction,
manufacturing, and wholesale trade industries[15] and GM's is
again significantly lower than the figure estimated by both
Argonne and the M3 FCM. However, use of employment growth
would again be expected to underestimate growth in VMT, since
employment grows more slowly than economic output due to
productivity improvements and heavy-duty VMT should more
closely follow the latter. For instance, if GM had chosen
growth in industry earnings (also included in DOC's
projections) instead of jobs as an indicator of future
heavy-duty travel, the new figure would be roughly 3.2
percent/year.[15] Thus, the FRM projections appear quite
-reasonable. .
This overall growth rate for heavy-duty VMT must then be
split between gasoline-powered and diesel-fueled vehicles
(HDGVs and HDDVs, respectively). The MOBILES .Fuel Consumption
Model determines this split using diesel sales penetration
rates developed along with MOBILE3,[10] the contents of which
were critiqued by vehicle and engine manufacturers and other
interested parties through a number of workshops.
3. Diesel Sales Projections
Manufacturers (primarily GM) recommended significantly
lower future light-duty diesel sales fractions than those
projected in the NPRM, suggesting 1995 model year diesel
penetrations of 5 percent and 7 percent for LDVs and LDTs,
respectively. These estimates compare to NPRM figures of 11.5
percent and 34 percent, respectively.
Future "light-duty diesel penetration is difficult to
predict, as the demand for diesels is very dependent upon
future oil prices and the availability of diesel engines which
satisfy consumer preferences. However, during the development
of the MOBILES heavy-duty conversion factors, manufacturers
(particularly GM) argued for substantial fuel economy
improvements well through the 1990's, indicating a belief that
fuel prices will indeed rise in the future calling for
continued improvements in fuel economy. Therefore, to remain
consistent with this position, growth in diesel penetration —
a fuel-saving technique — was also projected to occur. 'EPA
raised this issue at that time, indicating that substantial
vehicle-related fuel economy improvements must logically be
accompanied by increasing diesel usage. EPA accepted most of
these fuel economy improvements predicted by the manufacturers,
which lower heavy-duty emissions in the future without direct
emission control. Thus, to argue for low diesel penetrations
-------
now is quite inconsistent with GM's position just a year ago
and inconsistent with fuel economy improvements assumed in tne
derivation of the heavy-duty conversion factors.[10]
Therefore, model year diesel sales fractions used in the
FRM analyses are similar to those estimated in the NPRM
projections (post-1994 estimates of 11.5 and 34 percent for
LDVs and LDTs, respectively), except that pre-1995 estimates
have been reduced to reflect slowed growth (1990 projections of
5 percent and 15 percent, respectively). However, to identify
the impact of potentially lower diesel penetration on
particulate emissions, a sensitivity analysis will be performed
wherein the 1990 penetrations (5/15 percent) are held constant
through model year 2000 (results to be discussed in the final
section of this chapter).* A complete listing of the
light-duty model year diesel sales factions used in the FRM
analyses is provided in Table A-4 of the Appendix.
While current light-duty diesel penetration is relatively
low, particularly in light of GM's recent decision to withdraw
from the market, the 11.5 percent 1995 LDV penetration is still
realistic given that diesel penetration jumped from 0.3 percent
in 1977 to 6.0 percent in 1981 with only one domestic
manufacturer producing diesels. Given this fact, plus the
potential volatility of world oil prices, it is not difficult
to project a rapid increase in diesel sales if fuel prices were
to increase dramatically. Furthermore, in the development of
MOBILES and elsewhere, manufacturers have consistently
predicted a continued need in the next decade to improve the
fuel economy of their engines/vehicles, and EPA's diesel
penetration rates are not inconsistent with these forecasts.
In view of the current (1983) level of diesel penetration
into the LOT market — approximately 8 percent — and the fact
that the diesel fraction of LDT sales has been steadily
increasing since 1978, it is apparent that LDT diesels are a
growth market. Given this, GM's estimate of 7 percent for 1995
seems unrealistically low, particularly since GM supports the
need for future fuel economy improvements and does indeed
predict growth in diesel penetration of all other markets (LDV
and HDV classes). Therefore, 15 percent is a more realistic
lower limit for the sensitivity analysis, maintaining a best
estimate of 34 percent diesel penetration into the LDT market
by 1995.
NOx emission factors for gasoline and diesel LDVs and LDTs
are quite similar. Therefore, future NOx emissions are
not sensitive to light-duty diesel penetration.
-------
4-12
GM also commented on diesel penetration of selected
heavy-duty classes, recommending 1995 figures of 25 percent and
52 percent for heavy-duty Classes III-V and VI, respectively.
Although the NPRM analyses assumed slightly higher penetrations
for these classes, the use of MOBILES for the final rulemaking
projections implicitly assumes diesel fractions consistent with
the heavy-duty conversion factors analysis;[10] these figures,
also used as input to the MOBILES Fuel Consumption Model,
essentially are in agreement with GM's estimates (30 percent
and 53 percent for Classes III-V and VI, respectively). A
complete listing of the final heavy-duty diesel sales fractions
appears in Table A-5 of the Appendix.
4. Heavy-Duty Conversion Factors
A fourth issue — heavy-duty emission conversion factors
— has been addressed extensively in the MOBILES workshops and
documented in an August 1984 technical report.[10] No
commenter brought any new information to bear in this area. As
EPA has made known in the past,[16,17] MOBILES conversion
factors for both HDGVs and HDDVs are significantly lower than
those used in the NPRM analyses (based on MOBILE2.5). However,
GM's contention that even further fuel economy improvements
should have been incorporated (resulting in even lower
emissions) appears inconsistent with their projections of low
diesel penetration into the light-duty markets and slightly
lower projections for the heavy-duty market. Therefore, the
FRM analyses will continue to use the MOBILES conversion
factors. (The final MOBILES conversion factors are presented
in Table A-6 of the Appendix.[10])
5. Validity of Rollback Air Quality Models
The final issue common to both the NOx and particulate
analyses is the validity of the "rollback" approach to
predicting future air quality, where any change in emissions is
assumed to translate proportionately into a change in ambient
pollutant concentrations. In submitted comments, Ford (with
support from MVMA) estimated that only one-fifth to one-third
of- the change in emissions due to VMT growth, not the entire
change, should be applied to air quality projections; this
estimate is based on area source dispersion modelling conducted
by Ford.[18]
Investigation into Ford's urban analysis uncovered some
assumptions which could have biased the results of the study.
One, the traffic density (VMT/square mile) at the center of the
city was assumed to remain constant. While VMT growth at city
center is certainly rrore restricted than that at the outskirts,
this assumption allows absolutely no consideration for urban
redevelopment nor roadway construction or improvement.
-------
4-13
Fur therrrore, the assumed r.odel of VMT density forced post
growtn xn VMT to be applied to the outer edges of the original
urban area and to areas even beycnd the original city radius,
as the square of the city radius was increased in proportion to
assumed emission growth.
Two, the choice of location for the two air quality
monitors, when coupled with the above assumptions, also appears
to minimize the impact of motor vehicles. The first monitor,
located at city-center, would be primarily affected by the area
just upward of city-center, where VMT growth has been assumed
to be essentially zero. The second receptor, located 10 km
directly downwind of city-center, would also -be most affected
by emissions in areas again assumed to experience little
growth. Monitors not in line with the city-center, which were
not included in the study, would be expected to experience more
VMT growth than was assumed to be present in the more congested
areas, and would therefore be more likely to demonstrate the
impact of motor vehicle emissions.
An uncertainty present in Ford's urban dispersion
modelling is the selection of only one stability class.,
-"slightly unstable." As no information was given on the
characteristics of this and other classes, it is difficult to
assess the impact this choice had on the results.
EPA and others have used rollback modelling to project
future air quality since the mid-1970's, and EPA has long
approved its use in State Implementation Plans'for projecting
compliance. Validations of the rollback model as applied to
carbon monoxide and lead were included in Chapter 3 of the
Diesel Particulate Study;[5] and the figures presented there
show a strong correlation between emissions and ambient
concentrations over a decade. While dispersion modelling is
probably more accurate, it is not feasible in terms of expense
or time in a study such as this to evaluate every city using
dispersion modelling. Instead, a simpler approach, such as the
rollback model, must be used. Given the apparent bias and
uncertainties in the Ford study, it would be inappropriate to
discard or significantly adjust the rollback model here.
However, possible improvements to the rollback approach, such
as modified source discount factors, will be considered and
could be incorporated into future modelling efforts if merited.
6. Significance of the Air Quality Impact
Many comments were received concerning the significance of
the projected increases in urban concentrations of particulate
natter and N02 due to truck emissions. MVMA and DOE
questioned what portion of the future particulate ambient
levels' can be attributed to diesel trucks. The engine
-------
4-14
manufacturers also questioned whether the increases in NO^
levels warrant the standards that have been proposed. The
environmental interests, and most of the private citizens who
chose to comment, uniformly criticized EPA for, in their
impression, setting standards designed to hold emissions at
current levels and not attempting to achieve net reductions.
The standards that have been established for both light-
and heavy-duty truck NOx and heavy-duty diesel particulate have
been based upon requirements of Congress, which primarily focus
on technological feasibility and not only on environmental
"impact (the reader is referred to the Preamble to the final
rule). For example, with respect to the particulate standards,
the Act calls for the most stringent standards yielding "the
greatest degree of emission reduction achievable through the
application of technology which the Administrator determines
will be available...giving appropriate consideration to the
cost...and to noise, energy, and safety factors associated with
the application of such technology." Thus, the availability of
technology is the limiting factor — not satisfactory
environmental quality.
At the same time, the environmental impacts described in
the Draft RIA, and below in" Section III of this chapter,
clearly justify the need for the standards being promulgated.
Without these NOx standards, urban NOx levels would rise
significantly over current levels by the early 1990's in
low-altitude areas and even sooner in high-altitude are'as.
Even with these standards, growth in emissions is only being
delayed until the late 1990's at low altitude and there is
almost no delay of growth at high altitude. Nationwide NOx
emissions from all sources will also grow substantially by the
mid-1990's, even with s substantial reductions from these
standards. The case for the particulate standards is even
stronger, given the widespread noncompliance with the current
TSP NAAQS and that expected with the PMi 0 NAAQS (discussed
later). Thus, the arguments that the standards are either too
lenient or too strict based on environmental impact are not
valid.
B. Factors Specific to NOx
1. Stationary Sources
Although no comments were made pertaining to the
development of the stationary source inventories of NOx
emissions, nor their projected growth, these were reviewed in
light of what was discovered concerning the NEDS county
specific estimates of mobile source VMT. The methodology used
by NEDS to determine their inventories for stationary source
-------
4-15
t:0x were found to be acceptable. [ 19.20] Therefore, tne IIEDS
inventories (updated to 1932) used in the NPRM air quality
analysis will continue to be used here.
Tne growth rates associated with stationary source NOx
used in the NPRM were determined for EPA by EEA in 1979. These
were based upon certain population and industrial earnings
growth factors as determined by DOC/OBERS in 1977.[21] These
figures have been compared to those in the 1980 edition of
OBERS,[15] and the growth factors do not appear to have changed
significantly, so the same rates are being used here. A more
detailed review of this issue will be performed in the near
future as the National Acid Precipitation Assessment Program
begins releasing its projections. (The final stationary source
growth rates are presented in Table A-7 of the Appendix.)
2. NO; Ambient Design Values and Inclusion of
California
A second issue specific to the NOx analysis is the set of
N02 design values, or base-year ambient N02 concentrations,
used in the air quality projections for selected cities,
Commenters (Ford, MVMA) recommended the use of average
concentrations over a 3-year period to minimize the effect of
year-to-year fluctuations in monitored levels. This was in
fact already being done, as interim air quality analyses
conducted after completion of the NPRM analysis (early 1984)
were based on NOZ design values averaged over the period
1980-82. These design values are being updated once more for
this analysis, as design values for the years between 1981 .and
1983 are now available.[22] —
With the adoption of updated design values, the specific
cities that needed to be included in the NO: analysis (those
with concentrations at or above 0.035 ppm -- 66 percent of the
NO2 NAAQS of 0.053 ppm) are different from those cities
modelled in past studies. (Table A-8 of the Appendix lists the
cities included in past and current N02 analyses, along with
the NOZ design values used in the air quality projections.)
Also, as the monitoring period was updated to 1981-83, the base
v*»ar ehanaed to 1982 {the middle vear}: therefore. all base
year emissions inventories for mobile and stationary sources
(both discussed in previous paragraphs) used in the FRM
analysis are now calculated for calendar year 1982.
As Table A-8 shows, California cities were not included in
the NPRM NOx analysis, primarily because California vehicles
are certified under different (more stringent) standards
promulgated and enforced by the California Air Resources Board
(CARB) . However, CARB corroiented that many Federally-certified
(non-Ca'lif ornia) line-haul trucks cross over California state
-------
4-16
lines and contribute to NOx and particulate emissions in
California cities. Therefore, CARB feels that the impact of
Federal heavy-duty engine standards on California air quality
should be evaluated in the FRM. This is reasonable. Thus,
CARB's projections of NOx emissions for the South Coast Air
Quality Basin (SCAB), which includes the three California
cities shown in the last column of Table A-8, are presented in
the final section of this chapter.* (The inclusion of
California cities in the diesel particulate analysis was not an
issue, as all urban areas across the nation were modelled in
aggregate; in addition, air quality projections were included
for Los Angeles and San Diego in the DPS[5] and are included in
the aggregate results presented in both the NPRM and the FRM.)
3. NOx Emission Factors
Some commenters recommended that MOBILE3 NOx emission
rates be used instead of those in MOBILE2. This update was of
course made, beginning with interim analyses conducted while
-the NPRM was being reviewed in early 1984. [23} For use here,
the MOBILES inputs for post-1987 model year LDTs and HDEs were
updated to apply specifically to the following two scenarios:
1) • a "base case," which represents no further control of motor
vehicle NOx (2.3 g/mi and 10.7 g/BHP-hr for LDTs and HDEs,
respectively), and 2) a "controlled case," which evaluates the
effect of the final standards promulgated in this rulemaking
(1.2 and 1.7 g/mi for LOT, and LDT2, respectively, and 6.0
followed by 5.0 g/BHP-hr for HDEs). The emission rates used' in
the FRM analysis are summarized in Tables A-9 and A-10 for low
and high altitude areas, respectively; only those emission
rates and assumptions that are different from MOBILE3 are
provided.
It should be noted that the scenario designated as
baseline (2.3/10.7) in the FRM • analysis differs slightly from
the baseline scenario presented in the NPRM or in MOBILES. In
the proposal, future HDDV NOx emission rates were assumed to
remain at current levels (approximately 7.6 g/BHP-hr) even
though the standard was set at 10.7. In preparing the FRM
analysis, this previous assumption seemed unrealistic in light
of the pressure that a particulate standard would put on NOx
emissions, so the HDDV rates were instead adjusted upward
assuming manufacturers would design for the 10.7 standard once
they were sure it would remain at that level. Because -the
Because EPA's MOBILES program does not have the capability
to compute composite emission factors for California,
CARB's NOx emissions and air quality projections were
incorporated into the analysis.
-------
4-17
heavy-duty gasoline (HDGV) rates are currently well below the
standard and the particulate standards do not apply to these
vehicles, no adjustments to the previous assumptions for HDGVs
were r.ade.
4. Short-Term NO? Standard
The Natural Resources Defense Council (NRDC) commented
extensively on the need for a short-term (3-hour) N02
standard. The NAAQS for NOZ is currently under Agency
review. The Agency is currently involved in extensive research
concerning the potential need for such a standard. For the
time being, however, it is EPA's opinion that the current
annual' standard for N02 provides adequate protection against
both long- and short-term health effects associated with
NOZ . As the basis for the standards being promulgated is
technological feasibility, and not the limit of environmental
need, the existence of a short-term NO2 NAAQS should not
"affect this rulemaking, except to further justify the controls
being implemented.
5. Ozone and Acid Precipitation
Another issue specific to the NOx analysis is the effect
of NOx reductions on urban ozone and downwind sulfate
concentrations. GM (with support from several other commenters)
contends that a decrease in NOx emissions will cause urban
ozone and downwind sulfate levels to rise. NRDC, however,
disagreed with GM's view on ozone formation, citing various
sources who maintain that NOx control (as well as HC control)
is essential in the reduction of ozone levels. NRDC does
suggest that an increase in urban NOx emissions may lower ozone
levels locally (as GM contends), but it will also result in
increased ozone concentrations downwind of the higher NOx
emissions, merely delaying peak ozone formation.
The exact relationships between NOx and the other two
pollutants are rather complex and have been the subject of a
fair amount of controversy over the past decade. Numerous
factors play a role in these relationships, including
(specifically for ozone) the ratio of HC to NOZ ambient
concentrations, meteorological and topographical
characteristics of the area, spatial location of the NOx
reductions, and others. Therefore, the relationships could
differ from one urban area to another. In addition, existing
scientific studies of the NOx/sulfate and NOx/ozone
relationships are limited, and their results have not yet been
adequately reviewed or accepted by the scientific corr.r.unity.
An EPA-sponsored study of the NOx/ozone relationship is
currently underway; however, the results are not yet available
and, in any event, are unlikely to support net increases in NOx
emissions.
As will be shown in the final section of this chapter, the
NOx standards promulgated in the final rule will prevent
substantial growth in NOx emissions beyond current levels, but
-------
4-13
will not significantly decrease NOx emissions between 1982 and
the 1995-2000 time frame. Therefore, since a large reduction
in total NOx is not an issue here, no substantial increase in
ozone or downwind sulfate is suggested. Also, the possibility
of reducing ambient ozone or sulfate concentrations by allowing
NOx emissions to increase significantly is not now considered a
viable long-term option. To allow concentrations of one
dangerous pollutant (N02) to increase in hopes of lessening
other pollutant levels would not appear to be wise. Instead,
EPA will most likely address the need for further ozone and
sulfate control in the context of HC control strategies and
acid precipitation policy.
Several other comments were received concerning the
relationship between truck NOx emissions and acid
precipitation. The general comment from the manufacturers is
that controlling truck NOx emissions is an inappropriate way to
control acid precipitation, since it only represents a small
percentage of emissions producing acid precipitation. GM also
cites the fact that nitrate is much less acidifying than
sulfate.
Environmental groups (specifically NRDC) were, in their
words, appalled at the lack of any reference to acid
precipitation in the Draft RIA. They recognize that, overall,
SO2 has more importance in terms of acid precipitation, but
insist that NOx cannot be ignored. NRDC refers specifically, to
the Western U.S., where NOx contributes over half of the
acidity in precipitation, and to such seasonal events as the
spring snowmelt, where nitrates dominate the acidity.
There has been a great deal of controversy over acid rain
in recent years as to its causes and effects, primarily due to
the complexity of the issue and the lack of substantial
clear-cut data on the subject. Although knowledge of acid
precipitation is incomplete, it is clearly becoming a problem
over widespread areas of the country.
Although NOx emissions contribute only about a third of
all acid deposition in the east, [24] they may have a
disproportionately higher impact in terms of their effects.
For example, nitric acid tends to become concentrated in the
winter snowpack and is then released during the spring thaw,
creating episodic "hot spots" of acidity which unfortunately
tend to coincide with the spawning period for fish and the
beginning of new growth for plant life.[24]
In contrast to the east, NOx is the predominant acid rain
precursor in the western part of the United States. This is
due primarily to the use of low-sulfur coal in western
powerplants, which results in only 20 percent of annual U.S.
SO2 emissions being produced in the states west of the
Mississippi River.[24]
-------
4-19
Also, while SO.- is primarily emitted from stationary
sources, NOx production is a joint mobile source/stationary
source problem. As will be shown below in Section III, motor
vehicles are responsible for almost one-third of nationwide NOx
emissions. In the absence of further controls for LDTs and
HDEs, nationwide NOx emissions will increase by 23 percent
between 1982 and 2000. With these controls, emissions will
still increase 14 percent by 2000, but will be 8 percent lower
than uncontrolled levels, which represents a significant
reduction.
Thus, at this time, it cannot be concluded that motor
vehicle NOx controls have no effect on acid precipitation. Nor
can it be stated that such controls will play a large role in
acid precipitation control policy. Identification of the most
appropriate role for motor vehicle NOx control must wait for
the completion of the in-depth evaluations of the formation,
transport, and welfare effects of acid deposition which the
Agency has underway. However, as the health effects associated
with both current and future NOx emission levels justify the
need for these standards, this rulemaking need not wait for the
completion of the acid deposition studies.
6. Visibility Effects
• NRDC commented that NOx can play a part in visibility
degradation, either in the form of NO2 gas or nitrate
aerosols. They indicate that 31 percent of the light
extinction attributed to mobile sources in Denver in 1980 was
due to motor vehicle NOx emissions.
The effects of NO2 on visibility were examined in the
review of the NAAQS for nitrogen oxides.[25] Th6 conclusion by
EPA at that time was that, although N02 does have a
visibility impact, the improvement in visual air quality to be
gained by reducing N02 concentrations was uncertain at best.
Due to this uncertainty, NOx-related visibility impacts have
not been considered in this rulemaking. However, as the
standards being promulgated in this rulemaking will reduce
future N02 levels in the atmosphere from what they would have
been, to the extent N02 affects visibility, future visibility
should improve.
C. Factors Specific to Diesel Particulate
1. Health Effects
NRDC, along with other environmental groups, took issue
with how EPA characterized the health effects due to diesel
particulate matter. They agreed with the EPA's statement that
the cancer risk due to diesel particulate matter is
-------
4-20
"significant," but emphatically disagreed with EPA's assessment
of this risk as "small." NRDC also stated that "the proposal
notice makes no mention of the non-carcinogenic health threat
from fine particulate emissions."
On the other hand, GM and the American Trucking
Association (ATA) questioned the adverse health effects of
diesel particulate emissions. Citing studies by the British
Medical Research Council on London bus garage workers, the
conclusions of the National Research Council's Diesel Impact
Study committee and some of their own studies, GM concludes
that there is no definite evidence to implicate diesel
emissions as a "serious cancer hazard." ATA feels that since
"available evidence does not indicate that diesel exhaust
particles cause human cancers," any reference to such "should
"be removed from the record." They also question EPA's use of
relative potency analysis in determining the cancer risk
associated with diesel particulate matter.
"The non-carcinogenic effects of diesel particulate matter
were detailed in both the draft RIA and the DPS. [5] These
effects are compared to the effects for other inhalable
particulate matter (PM10, particulates less than • 10
micrometers in diameter), which, as opposed to TSP, appear to
be most directly related to adverse non-cancer health effects.
Based on the available data, no clear differences in
non-carcinogenic health effects between ambient PM|0 and f.ine
diesel particulate matter could be determined, though there is
some possibility that diesel particulate may be somewhat more
hazardous. Thus, when considering overall health impact, the
effect of diesel particulate control on PMio levels was used
as the primary indicator. As the commenters submitted no new
data to the contrary, this finding must stand.
The carcinogenic health effects associated with the diesel
particulate matter were also detailed extensively in the Draft
RIA and the DPS. [5] The studies on the London bus garage
workers were reviewed in the DPS and analyzed independently by
the EPA's Carcinogen Assessment Group. Flaws in the design of
these studies caused them to be disqualified from further
consideration in the DPS, and no new information has been
brought to light to change that determination. Another
epidemiological study is currently being conducted by Harvard
University to evaluate the possible effect of diesel exhaust in
U.S.. railroad workers. This study, referred to by NRDC, is
described in the DPS, and will be reviewed by EPA upon its
completion.
EPA did base its determination of the potential cancer
potency of diesel particulate upon a comparative potency
analysis that assumes that the relative results of lower animal
-------
4-21
testing can be extrapolated co hunans. While hur.an
eoidemiological data are definitely preferred, this approach is
not feasible until a reliable epidemiological study is
available. Until then, the relative potency analysis remains
the most reliable.
With respect to the estimated cancer risk, the approach
was taken to objectively state the risk and compare it to
others experienced by the populace. Given that the risk stated
is a lifetime risk for exposure to 1995 ambient levels of
diesel particulate, the risk does not stand out and call for
control beyond that which is technologically feasible for
diesels. However, at the same time, the risk is not negligible
and does support the need for some degree of control.
There was one additional comment on EPA's use of the
proposed PMio NAAQS to assess the effect of diesel
particulate emission control. MVMA feels that "it is
completely inappropriate for EPA to anticipate a PMio
standard, which has not been promulgated." They cite this as
an act of Mpre-judgment and a compromise of free ideas."
The proposed standards for PMio appear in the March 20,
1984 Federal Register, but have not yet been promulgated. Use
of this proposed NAAQS was thought to have provided the most
appropriate means of demonstrating the impact of diesel
particulate control on human health, as the change to PM]0
from TSP was proposed to more properly force control on those
particles affecting health. The diesel standards being
promulgated could just as easily have been based, on the current
TSP standards. Justification of the light-duty diesel
particulate standards was based on the TSP standards, and
noncompliance with the TSP NAAQS is projected tc be rrore
widespread than with the PM,0 standards.* Thus, use of the
proposed PMio standards provides another perspective from
which to assess the need for particulate control and does not
affect the result: diesel particulate control is justified
environmentally. The aspect affected is the precision to which
that need, and the effect of control, is identified.
2. Visibility Effects
Several comments were received pertaining to the
visibility impacts of diesel particulate matter. Based upon a
study of four cities, GM concluded that no significant impacts
Between 105 and 329 counties are projected to be in
non-attainment of the proposed PMio standard, compared
to 300-525 counties estimated to be in non-compliance with
the current TSP standard in the 1987-89 timeframe.[26]
-------
4-22
on visibility due to increased diesel particulate
concentrations will occur except under strict NOx controls (1.0
g/mi for LDV, 1.2 g/mi for LOT, and 4.0 g/BHP-hr for HDE) .
They appear to have set a 5 percent reduction in visibility as
the cutoff for "significant impact." NRDC, the Colorado
Department of Health, and several private citizens mentioned
their concern about visibility, especially in the Western U.S.
NRDC emphasized that the reductions in visibility given were
only averages, and that on many days the effect could be much
worse than indicated.
The methods by which the EPA estimates the visibility
impact due to diesel particulate matter are described in detail
in Chapter 4 of the DPS.[5] These estimates are highly
dependent upon the projections of diesel particulate
emissions. EPA and GM differ substantially in this respect as
is indicated by the analysis of other GM comments earlier in
this chapter. In the case of the four cities modelled: New
York City, Los Angeles, Washington, DC, and Denver, GM chose
not to project any VMT growth except for Denver. Also, a
fundamental difference lies in the value used for the critical
level of contrast against background required to determine
visibility. EPA used a value of 5 percent at airport sites .for
reasons described in the DPS. - If similar modelling techniques
are assumed (i.e., Beers' Law), GM's value is closer to 0.14
percent, which is well beyond the level of contrast discernable
by the human eye. Correcting for some of these differences and
considering the NOx standards being promulgated, the
differences in the resulting estimates of the visibility
impacts can be readily explained.
The projected reductions in visibility due to diesel
particulate are annual average reductions, and it is likely
that the effects will be greater on some days and less on
others. However, the level of sophistication of the model and
input data do not allow shorter term effects to be estimated
accurately.
3. Soiling Effects
A few comments were received concerning the impacts of
soiling due to diesel particulate matter. NRDC, in particular,
cites estimates of economic costs due to soiling ranging from
hundreds of millions to billions of dollars annually. EPA has
reviewed the scientific and economic literature pertaining to
soil'ing from particulate matter in general, and diesel
particulate matter specifically. The estimates of the benefits
from reduced soiling due to diesel particulate control shown in
Chapter 8 of the Draft RIA were in the same range as the
estimates quoted by NRDC. Therefore, there is general
concurrence on this issue.
-------
4-23
III. E.T.I ss ions/Ai r Quality Projections
Both in response to the comments analyzed in Section II
and co part of the ongoing process of re-evaluation and
improvement of EPA's modelling efforts, EPA has revised its
projections of future NOx and diesel particulate emissions and
air quality. Several of the input parameters to EPA's models
were revised with the adoption of MOBILES,[27] and as mentioned
earlier, the comments received on the emissions and air quality
model inputs were also given full consideration in the
development of final estimates for each parameter.
This final section of the chapter presents these revised
projections, based on EPA's current best estimates for each of
the various input parameters. Section A will deal with the NOx
projections, followed by a discussion of the diesel particulate-
analysis in Section B. In both analyses, the methodologies and
inputs are the same as those used in the NPRM analyses, except
for the input changes discussed in Section II (and detailed in
the Appendix). For information on the methodologies used, the
reader is referred to Section I of this Chapter and also to the
Draft RIA and the DPS.[5]
A. NOx Analysis
Projections of future NOx emissions and related air
quality both with and without the promulgated LOT and HDE
standards are presented below. First, the NOx analysis focuses
on emissions in key urban areas (low-a-ltitude, high-altitude,
and California), and then moves to projections of nationwide
NOx emissions. The third part of the NOx analysis deals with
the impact of future emissions on ambient NO, levels in the
urban areas of concern, and a final section offers EPA's
conclusions on the need for future NOx controls.
1. Emissions in Key Urban Areas
As mentioned above, the first part of the NOx analysis
focuses on the ten urban areas shown in Table A-8 of the
Appendix, consisting of eight , low-altitude and two
high-altitude cities. Also, CARB's projections for the three
California cities shown in the table (all located in the South
Coast Air Basin) are included in this discussion.
Table 4-2 presents base-year, and future "NOx emissions
inventories for the low- and high-altitude cities under two
future NOx standards scenarios. "Base case" represents no
further control of NOx, with a LDT standard of 2.3 g/rni and a
HDE standard of 10.7 g/BHP-hr. The "controlled case" refers to
the NOx standards being promulgated — 1.2 g/mi and 1.7 g/.-ni
for the LDTj and LDT2 classes, respectively, and HDE
-------
4-24
Table 4-2
Base-year and Future'Urban NOx Emissions*
(1000 tons/year)
Eight Non-California Low-Altitude Urban Areas
Source
'LDV
LOT
HDGV
HDDV
Others
Total
Source
LDV
LOT
HDGV
HDDV
Others
Total
1982
281
114
42
82
— 291
810
Two Non-Cali
1982
18.6
7.5
2.6
7.5
38.5
74.7
Base
203
126
39
146
352
866
f ornia
Base
18.6
11.4
2.8
13.3
47.3
93.4
1995
Controlled
203(0%)***
.105(17%)
35(10%)
84(42%)
352(0%)
779(10%)
High-Altitude
1995
Controlled
18.6(0%)*
9.4(18%)
2.5(11%)
7.7(42%)
47.3(0%)
85.5(8%)
Base
220
131
40
171
381
943
* *
2000
Controlled
220(0%)***
100(24%)
34(15%)
89(48%)
381(0%)
824(13%)
Urban Areas**
Base
** 20.3
12.1
3.0
15.6
51.5
102.5
2000
Controlled
20.3(0%)***
9.2(24%)
2.5(17%)
8.1(48%)
51.5(0%)
91.6(11%)
NOx emissions do not include those from stationary point
sources, due to limited air quality impact relative to
ground-level sources.
** Includes the eight low-altutude and two high altitude
SMSAs listed in Table A-9 (FRM column).
*** Numbers in parentheses represent reductions from base case.
-------
4-25
standards of 6.0 g/3H?-nr in 1958, followed by 5.0 g/E-iF-hr in
1S91. Stationary area and off-highway source NOx emissions are
included in the category "Others." In these urban projections,
staticr.ary point source emissions are not included because of
their relatively low air quality impact per ton compared to
that of ground-level sources.
As shown, total baseline NOx emissions in the eight
low-altitude urban areas are expected to grow by seven percent
between 1982 and 1995, with an overall increase of 16 percent
by the year 2000. As in the NPRM, the largest increase is
projected for the HDDV class, with year 2000 emissions more
than double the 1982 levels. LDT emissions increase by
approximately 15 percent, while HDGV and LDV emissions decrease
without further control. (Shown graphically in Figure 4-1.)
The effect of the final standards for LDT and HDE NOx
emissions in these eight low-altitude areas is also evident
from the projections in Table 4-2. As shown, controlled NOx
emissions are estimated to be approximately 10 percent lower
than the base case in 1995, and 13 percent lower in the year
2000. These reductions due to stricter LDT and HDE NOx control
result in total NOx emissions (including those from stationary
area and off-highway sources) staying fairly constant through
the year 2000. Total emissions decrease by 4 percent in 1995
relative to 1982, and are roughly 2 percent higher than base
year in 2000, with motor vehicle emissions 18 percent lower in
1995 and 15 percent lower in 2000 (with respect to 1982
emissions). (See Figure 4-2.)
As shown in the bottom portion of Table 4-2, future
emissions growth in the high-altitude areas is projected to be
much greater than in the low-altitude cities. The difference
is not VMT growth, as the same national average rates were used
for both low and high-altitude areas; instead, growth is higher
because the 1.0 g/mi NOx standard on 1981 and later model yea-r
cars (LDVs) and the 2.3 g/mi standard on LDTs (beginning in
1979) did not have as great an impact on high-altitude
emissions as they did at low altitude. This is due to the fact
that pre-control emission rates , for LDVs and LDTs in
high-altitude areas were lower than those in low altitudes but
controlled levels are about the same. Therefore, the smaller
impact of existing light-duty controls on high-altitude
vehicles does not outweigh the future VMT growth, as it does in
low-altitude areas. This is shown in Table 4-2, where
base-case LDV emissions show a decrease between 1982 and 1995
in the low- altitude areas, but stay the same in high altitudes.
Overall, total baseline NOx emissions in the two
high-altitude areas are projected to grow by 25 percent between
1982 and 1995 (shown in Figure 4-3), compared to 7 percent in
-------
Figure 4-1
900-
800-
•D
S 700-
J eoo-l
o
O 500-
o
£ 400-
•^ 300-
LJ
X 200-
"Z.
100-
0
NOx Emissions Inventory for Eight Urban Areas
; Low Altitude
. Base Scenario
29li
82
810
519
146
866
514
1982
171
1995
2000
943
TOTAL
562
MOBILE SOURCES
N)
Legend
OTHER SOURCES
CD MODE
C3HDGE
Z2 LOV
-------
i. -.'Ore: •<-
900-i
800-
S 700-
| 600-)
O
O 500-
C 40°"
.£ 300-
UJ
X 200-
Z
100-
0
NOx Emissions Inventory for Eight Urban Areas
Low Altitude
1.2/1.^6.0/5.0 Scenario
82
810
519
84
779
427
89
824
TOTAL
ts>
443
MOBILE SOURCES
Legei
G23 OTHER SOu
a HDDE
OHDGE
E3 LOT
EZ1 LDV
1982
1995
2000
-------
110-i
100-
*- 90-
o '
0)
c
o
80-
70-
60-
£ 50-
*S 40"
l§ 30H
X
§ .20-
10-
0
Figure 4-3 . • |!
Emissions Inventory for Two Urban Areas
High Altitude f
Base Scenario
74.7
36.2
13.3
93.4
46.1
15.6
102.5
TOTAL
51.0
MOBILE SOURCES
4ft-
I
K>
CO
Legend
OS OTHER SOURCES
OHDDE
CSHDGE
C3LDT
£2 LDV
1982
1995
2000
-------
4-29
lev: altitudes. However, the promulgated LDT and HDE NOx
standards have basically the sarr.e effect on 1995 ar.d 2COO
emissions in both altitudes. This is expected because, by that
time, . as mentioned above, the baseline (2.3/10.7 standards)
emission rates in low and high altitudes are quite similar.
But even with the stricter control on LDTs and HDEs,
high-altitude emissions are expected to grow by 14 percent
between 1982 and 1995, with a 23 percent increase by the year
2000 (see Figure 4-4). These figures are quite large compared
to the relatively small changes from base-year levels projected
to occur in low-altitude areas with the added control.
Projections of future NOx emissions for- the South Coast
Air Basin (the Los Angeles area) were provided by the
California Air Resources Board (GARB) and are presented in
Table 4-3. GARB examined three NOx standards scenarios for
Federal line-haul (Class VIIIB) diesel trucks: 10.7 g/BHP-hr
(no further control), 6.0 g/BHP-hr in 1988, and finally the 6.0
standard followed by 4.0 g/BHP-hr in 1991.* Although the
Federal standard of 5.0 being promulgated in the final rule was
not specifically examined by CARB, sufficient data was provided
to interpolate between scenarios. All scenarios assume that
only Federal line-haul trucks will cross into California (i.e.;
none of the lighter classes).
As Table 4-3 shows, total NOx emissions (including
stationary point sources) in the SCAB are projected to be lower
than current levels in the year 2000, regardless of Federal
control. However, based on the California State Implementation
Plan (SIP), total SCAB emissions must be at or below 895
tons/day in order for the cities in the basin to be in
attainment of the NOZ NAAQS. CARB projects attainment to be
achieved sometime in the late 1330s, but projects
non-attainment by 2000 due to growth unless Federal (and thus
California) engines are certified at 4.0 g/BHP-hr. However,
(though not modelled by CARB) a Federal standard of 5.0
g/BHP-hr may result in only marginal non-attainment, based on
evaluation of the relative emission totals presented in Table
4-3.
2. Nationwide Emissions
In addition to evaluating the effect of the final
standards on NOx emissions in these specific low-altitude,
high-altitude and California urban areas, the impact on total
Due to provisions of California's waiver from Federal
standards, this Federal truck scenario also assumes the
reduction of California's standard from 5.1 to 4.0
g/BHP-hr.
-------
v_
D
-------
4-31
Tacle 4-3
South Coast Air Basin (SCAB)
NOx Projections (tons/day)
Source
LDV
LDT
HDGV
HDDV
2000 Federal
1983
327.9
94.5
44.2
134.0
113.0
200.7
98.9
1013.2
10.7
227.2
60.0
36.1
169.6
146.5
205.6
101.3
946.3
6.0
227.2
60.0
36.1
137.1
146.5
205.6
101.3
913.8
HDE Std. Scenarios
6.0/5.0-
227.2
60.0
3-6.1
130.5
146.5
205.6
101.3
907.2
6.0/4.0
227.2
60.0
36.1
105.3
146.5
205.6
101.3
882.0
Stat. Area
Total**
Degree Above
NAAQS Attainment
Level (%)*** 13 6.2 1 -1
* The 6.0/5.0 Federal scenario was not examined by CARB, but
- was estimated by EPA based on CARB's data; represents very
• rrisrgir.3l nor.attainrrent of NAAQS-
** Totals are up to 2 percent greater than those provided by
CARS, due to round-off error in recombining source
categories.
*** The California SIP estimate is that NOx emission levels at
or below approximately 895 tons per day are necessary for
SCAB attainment of the NO2 NAAQS. Based on this, the
6.0/4.0 Federal standard, which would be accompanied by a
reduction of the California standard from 5.1 to 4.0,
allows the SCAB to stay in attainment in 2000. (Initial
attainment is projected for the late 1980's, regardless of
Federal control.)
Source: California Air Resources Board, Mike Sheible,
January 22, 1985, phone conversation.
-------
4-32
nationwide NOx emissions was also determined. This larger
scale analysis can be especially useful in evaluating the
secondary effects of NOx contro-1, such as acid rain formation.
Because nationwide projections were not included in the NPRM, a
brief explanation of the methodology used is in order.
Projections for the nation (48 continental states) are
made using base-year inventories from the National Emissions
Data System (NEDS).*[1] Motor vehicle inventories are adjusted
for future VMT growth and emission control using nationwide
average VMT growth rates from the MOBILES Fuel Consumption
Model (shown in Table A-7) and MOBILE3 emission factor ratios
for the various standard scenarios. Current emissions from
other sources are adjusted using assumptions also shown in
Table A-7.[21,28] In this nationwide analysis, stationary
point sources are included due to the larger scale regional
concerns usually associated with secondary NOx effects.
These nationwide NOx projections are shown in Table 4-4
and in Figures 4-5 and 4-6; these "base" and "controlled" cases
refer to the same standards scenarios described earlier. As
shown, without further LOT and HDE control, total nationwide
NOx emissions are projected to grow by 13 percent between 1982
and 1995, with a 23 percent increase by the year 2000.
However, with the final LOT and HDE standards in place, growth
during the same periods is estimated to be 6 and 14 percent,
respectively, or an overall reduction of 6-8 percent from
future uncontrolled emissions.
3. Air Quality
Using the rollback model and input data described in the
Draft RIA and Section II above, the effect of the final NOx
standards on ambient N02 concentrations was evaluated for the
eight low-altitude and two high-altitude urban areas mentioned
earlier. Table 4-5 presents the results of this evaluation,
comparing projected N02 NAAQS attainment status under both
the promulgated standards and the baseline case. Because the
rollback approach was used, the percent change in ambient NOZ
concentration tracks the change in NOx emissions (excluding
point sources), which have already been described above.
Because the NEDS weaknesses exist primarily in ' the
apportionment of VMT to individual counties and do not
apply to statewide totals, the methodologies used to
calculate nationwide NOx inventories are appropriate for
use in this part of the analysis.
-------
4-33
Table 4-4
Total Nationwide NOx Emissions
(1000 tens/year)
1995
2000
Source
LDV
LOT
HDGV
-- -" HDDV
On-Highway
Vehicles
~-~-- (subtotal)
Stationary Area
Combustion
Off-Highway
Stationary
Point
.->.; Total . .
1982
3,082
1,134
466
2,256
6,938
241
3,013
1,941
10,847
22,981
Base
2,204
1,249
415
3,296
7,164
241
3,342
2,677
12,583
26,007
Controlled
2,204( 0%)*
1,038(17%)
368(11%)
1,903(42%)
5,513(23%)
241( 0%)
3,342( 0%)
2,677( 0%)
12, 583 ( 0%)
24,356( 6%)
Base
2,422
1,302
-420
3,699
7,843
241
3,478
3029
13,776
28,367
Controlled
2,422( 0%)*
989(24%)
357(15%)
1,925(48%)
5,693(27%)
241( 0%)
3,478( 0%)'
3,029( 0%)
13,776( 0%)
26,217( 8%)
Figures in parentheses indicate reductions from base case
-------
Figure 4-5
TOTAL NOx EMISSIONS - NATIONWIDE
Base Scenario: 2.3/10.7
30000-1
O
0)
25000H
20000H
O
O
O
O 15000
(/)
c
.2
'in
.82 10000
LJ
x
Z 5000H
0
28,367
TOTAL
J9R2
2000
.f
I
Legend
STAT PT
KH OFF-HWY
C3 COMBUST
£3 STAT AREA
CD HDDE
ED HDGE
a LOT
LDV
-------
Figure 4-
O
o>
O
O
O
oo
c
LJ
X
O
TOTAL NOx EMISSIONS - NATIONWIDE
Controlled Scenario: 1.2/1.7; 6.0/5.0
30000-1
25000-
20000-
15000 H
£ 10000-
5000 H
0
22,981
TOTAL
2256
1982
g^
/:
;>:>
«5«3g
5^
?368!
Ft038|
24356
TOTAL
26,217
TOTAL
5A92?>
1925
;357i
|909l
/////
//2422\
/////<
2000
OJ
Legend
STAT PT
EH OFF-HWY
CS COMBUST
BS3 STAT AREA
CD HDDE
E3 HDGE
C3 LOT
CZ3 LDV
-------
4-36
Table 4-5
Average Percent Change in NOx Emissions and
Ambient NO> Concentrations from the Base Year (1982)*
Eioht Low-Altitude Areas**
Base Case:
(2.3/10.7)
Controlled Case:
(1.2/1.7; 6.0/5.0)
Base Case:
(2.3/10.7)
Controlled Case:
(1.2/1.7; 6.0/5.0)
1990 1995 2QOO
-1 +6 +16
-6 -5 ' + 1
Two High-Altitude Areas**
1990
+ 13
+ 9
1995
+ 26
+ 15
2000
+ 39
+ 24
* Stationary point sources are not included in the emission
reductions.
** Negative value denotes a decrease; positive value denojtes
an increase.
-------
•1-37
Table 4-6 estirates the number cf Standard Metropolitan
Statistical Areas (SMSAs), or urban areas, projected to oe
above the ambient NO, standard of 0.053 ppm in several
pro^ec'ion years. It should be noted that actual number of
non-attainment areas shown is not to be taken as absolute, as
projections of this type are difficult to make. Rather, the
relative number of exceedances is more appropriate as a means
of evaluating the relative impact of a particular control
scenario. As shown, two of the three non-California areas fall
into attainment with the final standards in place, with the
three California cities predicted to be in only marginal
non-attainment in the year 2000.
4. Conclusions --
It is against the background of the above projections that
EPA must evaluate the comments by manufacturers that there is
insufficient need for NOx control to justify the proposed
standards for light-duty trucks and heavy-duty engines. Even
with the revised input data that project lower future
emissions, overall growth in future NOx is still projected to
be significant for both the nation as a whole and for the urban
areas of concern. The same basic need for further NOx control
demonstrated in the proposal still exists, and current action
is necessary if future problems are to be dealt with
effectively.
The statutory provisions of Section 202(a)(3)(E) allowing
EPA to relax the NOx standards based upon air quality
considerations place a positive burden on the Agency to
substantiate a lack of need for more stringent levels. Based
upon its projections of future emissions and their telationship
to both the attainment of the National Ambient Air Quality
Standard and to other actual or potential secondary impacts,
EPA finds it impossible to make such a statement at this time.
Therefore, the standards promulgated in the final rule have
been developed under the provisions of Section
202(a)(3)(B)-(D), which provide for setting standards based
upon those levels which do not increase cost or decrease fuel
economy to an excessive and unreasonable degree.
B. Diesel Particulate Analysis
Revised projections of diesel particulate emissions and
related impacts are presented in the following paragraphs. The
analysis begins with urban emissions projections both with and
without the promulgated HDE control, followed by a discussion
of the impact of these diesel particulate emissions on urban
air quality. The final section deals with the health and
welfare impacts of diesel particulate exposure, including both
non-cancer and carcinogenic health effects, visibility
-------
4-33
Table 4-6
Number of SMSAs Projected to Exceed the N02
Ambient Air Quality Standard
1984 1990 1995 2000
Base Case: (2.3/10.7)
Low Altitude 0011
High Altitude 0122
California • I* 0_ Q_ 3_
Total 1136
Controlled Case: (1.2/1.7; 6.0/5.0)
Low Altitude 0000
High Altitude 0111
California ~ ' ' 1* 0 - 0 3_
Total 1 1 1 4
Los Angeles is the only SMSA currently in non-attainment
of the NOj NAAQS.
-------
4-39
reduction, and soiling. Unless specified, the analyses
presented below utilize the rr.ethodology outlined in the Draft
RIA, as modified in Section II above.
1. Urban Emissions
Unlike NOx, diesel particulate is modelled for urban areas
across the nation in aggregate, without focus on particular
cities. This is done because violation of the NAAQS for
particulate is more widespread than it is for NOx. The final
diesel particulate emissions projections are presented in Table
4-7. The two future scenarios shown differ only in the HDDV
standards assumed; for light-duty diesels, the standards that
are currently set to come into effect with the 1987 model year
— 0.20 and 0.26 g/mi for LDDVs and LDDTs, respectively — are
assumed. The "base" scenario represents no further control of
HDDV particulate emissions, assuming uncontrolled emissions at
0.70 grams per brake-horsepower-hour (g/BHP-hr). The
"controlled" case is based on the HDDV standards being
promulgated in this rulemaking — 0.60 in 1988, followed by
0.25 in 1991.and 0.10 g/BHP-hr in 1994. (Urban diesel buses
will be subject to the 0.10 g/BHP-hr standard in 1991.)
As Table 4-7 indicates, urban diesel particulate emissions
are projected to grow to twice the current levels by the year
2000 if no further HDDV controls are imposed (shown graphically
in Figure 4-7). It is this HDDV category that makes up the
majority of the total emissions, representing 84 percent in
1984 and 63 percent of the total in 2000. (This decrease in
heavy-duty share occurs as the diesel penetration of the
light-duty market increases.) Table 4-7 also includes a
breakdown by class of the HDDV emissions, which shews that
line-haul (Class VIIIB) diesels make up almost half of total
HDDV emissions in 2000.
The effect of HDDV and urban bus control is significant,
with the combined 1988/91/94 standards bringing about an
estimated 46 percent decrease from the base (uncontrolled) case
in the year 2000. This level of control essentially prevents
significant growth beyond current levels, with about an 11
percent increase projected between 1984 and 2000 (see Figure
4-8).
The more stringent control (0.10 g/BHP-hr standard) of
urban buses, beginning with 199.1 models, and of other
heavy-duty classes in 1994 is a substantial portion of this
overall impact on emissions by the year 2000. The -0.10
g/BHP-hr standard accounts for 23 percent of the reduction in
emissions from uncontrolled levels. From another perspective,
if the 1994 0.10 standard were not implemented and the 0.25
standard simply continued on through 2000 for both buses and
-------
4-40
Table 4-7
Base-year and Future Urban Diesel ParticuLate
Bnissions (cons/year)* ___^_
1995 HDEV Scenarios
2000 HDCV Scenarios
Vehicle
Classes
1984
Levels
Base
(0.70)
Controlled
(0.60/.25/.10)
Base
(0.70)
Controlled
(0.60/.25/.10)
LDDV . 5,699 (11%)** 13,392 (15%) 13,392 (23%) 19,700 (18%) 19,700 (33%)
LDDT 2,492 (5%) 13,072 (15%) 13,072 (23%) 20,713 (19%) 20,713 (35%)
HDDV - 45,013 (84%) 61,485 (70%) 30,767 (54%) 68,528 (63%) 18,903 (32%)
Total 53,209(100%) 87,949(100%) 57,231(100%) 108,941(100%) 59,316(100%)
Breakdown of HDIV Bnissions (tons/year)*
Vehicle
Classes
2B-8A
83
Buses
Total
1984
Levels
15,427(34%)**
21,811(49%)
7,780(17%)
45,018(100%)
1995 HDCV
Base
(0.70)
24,062(39%)
26,710(44%)
10,713(17%)
61,435(100%)
Scenarios
Controlled
(0.60/.25/.10)
12,355(40%)
13,790(45%)
4,622(15%)
30,767(100%)
2000 HDDV
Base
(0.70)
27,343(40%)
28,909(42%)
12,276(18%)
68,528(100%)
Scenarios
Controlled
(0.60/.25/.10)
7,541(40%) '
8,146(43%)
3,216(17%)
18,903(100%)
* "Best estimate" diesel sales fractions, shown in Table A-5, are assumed.
** Figures in parentheses indicate percent of total.
-------
Figure 4-
Urban Diesel Particulate Emissions
Base Scenario
o
o
110-1
100-
90-
J 80H
8 7
-------
Figure 4-8
Urban Diesel Particulate Emissions
Controlled Scenario
HDDE: .6/-25/-1
BUSES: A/A
jr-
Nl
Legend
CD BUSES
C3 HDDE
C3 LDDT
IZ2J LDDV
-------
4-43
trucks, total diesel particulars erissions in the year 2000
would be approximately 33 percent higher than in 1932; nowever,
with the final more stringent standards, growth during this
period, is constrained to an estimated 11 percent.
The emissions projections presented in Table 4-7 are based
upon EPA's best estimates for the various input parameters;
however, because of the difficulty in projecting future diesel
penetration into the light-duty markets, a sensitivity analysis
was performed. Instead of assuming that light-duty diesel
production will continue to grow through 1995 (as in the "Best
Estimate" analysis), another case was examined wherein 1990
levels of 5 percent and 15 percent diesel penetration of the
LDV and LOT markets, respectively, was assumed to continue
through the year 2000. (Best estimates of heavy-duty diesel
penetration — less difficult to predict — were used in both
cases.) Results of the sensitivity analysis are presented in
Table 4-8.
As indicated, the use of the lower future diesel
penetrations results in a 29-30 percent decrease in light-duty
emissions in 1995 and a 47-50 percent decrease in the year
2000, in comparison to best estimate projections for the same
two years. With respect to total diesel particulate emissions
under the "Low Penetration" scenario, assuming no further
control, growth between 1984 and 2000 would still be
significant at 68 percent (compared to 105 percent assuming
"Best Estimate Penetration"). With imposition of the
1988/91/94 standards on HDDVs, assuming low diesel penetration,
year 2000 emissions would be approximately 25 percent lower
than current levels (compared to the 11 percent increase over
current levels projected using best estimates of light-duty
diesel penetration).
2. Air Quality
The impact of growth in diesel particulate emissions on
urban air quality is significant, as shown in Table 4-9.
Current ambient diesel particulate concentrations in large
cities are projected to grow from an average of 1-3 ug/m3 to
levels of 3-7 ug/mj by the year 2000 with no further control
on HDDVs (using best estimate assumptions). With the standards
promulgated in the final rule, diesel particulate
concentrations in large cities will be reduced to 1.5-4 ug/mj
(best estimates), a reduction to almost half of baseline
concentrations.
3. Health and Welfare Effects
As discussed in the Draft RIA and the DPS, [5] exposure to
diesel ' particulate emissions has an impact on these four
-------
4-44
Table 4-8
Sensitivity Analysis of Light-Duty Diesel Penetration
Urban Diesel Particulate Emissions (1000 tons/year;
1995
Vehicle
Class
LDDV
LDDT
HDDV**
TOTAL ' '"
1984
Levels
5.7
2.5
45.0
53.2 * -
Best
Estimate
13.4
13.1
61.5
~ 88.0
Low
Growth
9.4
9.3
61.5
- 80.2
2000
Best
Estimate
19.7
20.7
68.6
109.0
Low
Growth
9.9
11. Q
68.6
89.5
Standards scenario: no further HDDV control (0.7
"g/BHP-hr). LDDV and LDDT emissions do 'not change with
heavy-duty control scenario.
HDDV class includes buses.
-------
4-43
•Sable 4-9
Effect of Diesel Particulate Control on Urban Mr Quality*
Total Diesel Particulate Concentration (ug/ra^)
1995** 2000**
Ambient Urban Concentrations***
City Population: -
- Greater than 1,000,000
- 500,000 - 1,000,000
250,000 - 500,000
100,000 - 250,000
annual Average Exposure to U.S.
TOTAL
Microscale Concentrations
Roadway Tunnel:
Typical
Severe
Street Canyon:
Typical
Severe ,
On Expressway:
Typical
Severe
1984
1.3-3.0
0.8-2.0
1.0-1.6
0.7-1.7
Urban
2.4
63
159
2
16
7
28
Base
2.3-5.5
1.5-3.6
1.8-3.0
1.2-3.2
Dwellers
4.4
91
231
3
. 23
11
41
Controlled
1.5-3.6
. .1.0-2.4
1.2-2.0
0.8-2.1
2.9
60
152
2
15 '
7
27
Base
2.9-6.8
2.0-4.6
2.2-3.7
1.5-4.0
~ - 5.5
105
266
4
.26
11
48
Controlled
1.6-3.7
1.1-2.5
1.2-2.0
0.8-2.2 '
3.0
57
145
2
..14
6
26
Beside Expressway
* Based on best-estimate projections.
** Control effectiveness is approximately 35% in 1995 and 46% in 2000.
*** Ranges are average values plus and minus one standard deviation.
-------
4-46
areas: 1) non-cancer health effects, 2) carcinogenic health
effects, 3) visibility, and 4) soiling.
a . Non-Cancer Health Effects
Particulate matter in general has long been regarded as
hazardous to human health. EPA recognized this danger and
established an NAAQS for total suspended particulate (TSP) as
early as 1971. As discussed in Section II, EPA has proposed an
ambient standard that will focus on inhalable particles (i.e.,
those with diameters of 10 microns or less (PMi0))» because
it is this fraction that appears to be responsible for most- of
the human health effects associated with TSP.
As mentioned earlier, diesel-particulates fall easily into
the PMto category, as the majority are classified as fine
particulate (less than 2.5 microns in diameter). Although a
large body of data has been developed regarding the health
effects of inhalable particulate matter, research limited
specifically to diesel particulate is relatively new and
somewhat inconclusive. An analysis of the available data
indicates that, until more is known, diesel particulate
generally should be regarded as being equivalent to other forms
of inhalable particulate matter in terms of the hazard it
presents to human health, although there is a possibility it
may be somewhat more hazardous. [5] It should be pointed out,
however, that even if regarded as posing the same hazard,
diesel particulate is emitted directly into the breathing zone,
rather than from tall stacks that would promote dispersion.
Thus, the potential for human exposure is maximized.
Two basic concerns exist with respect to the health risk
posed by inhalable particulate in general. First, inhalable
particulates are small enough so that they are not as readily
prevented by the natural body defenses from reaching the lower
respiratory tract, as would coarser particles. Fine
particulate matter can penetrate to the alveoli, or deepest
recesses of the lungs, where the oxygen/carbon dioxide exchange
takes place with the circulatory system.[29] The body requires
months or years to clear foreign matter from the alveolar
region, as opposed to hours or days to clear the -upper
respiratory system. The second concern is that inhalable
particulate may be composed of toxic materials or may have
hazardous materials adsorbed onto its surface.
The most obvious non-cancer health effect of an inhalable
particulate, such as that produced by diesels, is injury to the
surfaces of the respiratory system, which could result in
reduced lung function, bronchitis or chronic respiratory
symptoms. The hazardous chemicals that may be associated with
particulate matter (e.g., organic compounds, lead, antimony,
-------
4-47
etc.) can either react witn lung tissue or be transported to
other parts of the body by the circulatory system. Particuiate
matter may also weaken the resistance of the body to infection
and vnere are indications that it reacts adversely in
conjunction with other atmospheric pollutants. For example,
studies in London, New York, Buffalo, and Nashville have found
an increase in the mortality rate, especially among older
persons, when high particulate levels were accompanied by high
sulfur dioxide levels.[30]
From the above discussion, it is clear that inhalable
particulate matter (PM10) has been linked directly with a
myriad of adverse non-cancer health effects, a-nd it is based on
this information that EPA has proposed the NAAQS for PMi0.
Also, diesel particles are all inhalable particulate and,
therefore, can potentially represent the same concern. "This
relationship can be used to assess the overall benefits of
controlling HDDV diesel particulate.
As stated in the Draft RIA for the PM,0 NAAQS, 105-329
counties are projected to be in non-attainment of the range of
primary PMio standards being considered for 1989.[26] Even
after reasonable non-mobile source emission controls are
implemented, numerous violations of the NAAQS are still
projected to occur. As shown in Table 4-9, if no further HDDV
controls were implemented, annual average exposure to diesel
particulate for urban dwellers would be at a level of
approximately 5.5 ug/m3 in the year 2000, or about 10 percent
of the suggested PMio NAAQS. Promulgation of the HDDV
standards is projected to reduce this exposure to about 3.0
ug/m3, therefore playing an important role in reducing urban
PMjo exposure. Furthermore, ths resulting reduction in
diesel particulate emissions within urban areas that continue
to violate the suggested PM,0 NAAQS will directly reduce the
non-cancer health effects associated with inhalable
particulates in general.
b. Carcinogenic Health Effects
A number of studies have concluded that exposure to diesel
particulate probably poses an additional risk of acquiring lung
cancer. EPA surveyed these studies and developed
scenario-specific risk factors for lung cancer incidence,
taking into account the relative reduction of compounds
producing the cancer-risk with respect to reductions in total
diesel particulate.[5] Table 4-10 shows the resultant cancer
risk estimates associated with diesel particulate for both- the
base-case and the controlled-case scenarios along with
estimated risks from other known carcinogens, shown for
purposes of comparison.
-------
4-48
Table 4-10
Comparison of Risks from Various Sources
Sources of Risk
Estimated Annual Risk
(risk/person-year)
Commonplace Risks of Death
Motor Vehicle Accident
Drowning
Burns
Tornados, Floods, Light-
ning, Hurricanes, etc.
Risks of Cancer Incidence
Diesel Particulate (1995):
Base Scenario 1.2
Controlled Scenario 0.8
Diesel Particulate (2000):
Base Scenario 1.5
Controlled Scenario 0.8
Natural Background Radi-
ation (sea level)
Average Diagnostic Medical
X-Rays in the U.S.
Frequent Airline Passenger
(4 hours per week
flying)
Four Tablespoons Peanut
Butter Per Day (due to
presence of aflatoxin)
Ethylene Dibromide
One 12-Ounce Diet
Drink Per Day
Arsenic
Miami or New Orleans
Drinking Water (due
to presence of chloroform)
Lung Cancers:
For Smokers Due to
Smoking
For General Population
222.
26.
21.
2.
x 10-
x 10'
x 10-
x 10"
20.
20.
10.
8.
4.
2.
1.
1.
419.
73.
0
0
0
0
6
6
t
t
0
0
0
0
2
6
7
0
0
9
X
X
X
X
_
-
_
-
X
X
X
X
X
X
X
X
X
X
10"
10"
ID'6
10-'
6.2 x 10-'
4.1 x 10'*
7.7 x 10"'
4.2 x 10"'
10"
10"
10-'
10"
ID'6
10"
10"
10"
10"
10-'
Exposed
Population
Entire U.S.
Entire U.S.
Entire U.S.
Entire U.S.
Urban U.S.
Urban U.S
Entire U.S.
Widespread
Limited
Fairly
Widespread
Widespread
Widespread
1% of U.S.
Southern
U.S., Urban
Entire U.S.
Due to Causes Other
Than Smoking
-------
4-49
The data indicate tnat wniie the ris* of contracting lung
cancer is greatest from smoking, exposure to diesel particulate
T.ay represent a significant portion of all non-smoking-related
lung cancer. The upper limit of the uncontrolled (base)
scenario in 2000 would represent almost eight individuals in a
million, or 10 percent of all non-smoking-related lung cancer
in the U.S. The lower limit still represents over one in a
million individuals, which has been used in the past by
regulatory agencies as a cut-off point for determining the need
for control. Thus, as indicated in the NPRM, Table 4-10 shows
that a relatively small but significant cancer risk may be
attributable to diesel particulate exposures. The promulgated
HDDV controls are estimated to reduce this' risk by almost
one-half in the year 2000.
c. Visibility Effects
. . Reduced visibility is one of the more readily apparent
effects of diesel particulate. Because diesel particles are of
a diameter most effective in scattering light and their 65-80
percent carbon content produces a high degree of light
absorption, visibility reduction results.
Table 4-11 presents the estimated visibility impacts of
the base- and controlled-case scenarios in terms of the average
percent reduction due to diesel particulates in 1995 and 2000
urban visibility from early 1970's levels. As shown, in the
absence of controls, increases in diesel particulate levels
will result in reduced visibility, ranging from a 22 percent
reduction in largest cities, to 4-9 percent decreases for less
populous urban areas in the year 2000. HDDV control is
projected to euu these visibility reductions to 12 percent in
the largest cities, and to 2-5 percent in smaller urban areas.
The controlled-case scenario thus offers a 2-10 percent
improvement in visibility over the base-case scenario,
depending on the size of the city. The lower limit of this
impact (i.e., the effect for smaller cities), may not be
perceptible. However, the effect for large cities would show a
noticeable improvement in visibility. The promulgated
standards, therefore, will provide, an overall benefit that
would be most apparent in the areas where it was most needed.
d. Soiling Effects
In a review of the scientific -literature, the DPS[5] found
some evidence suggesting that because of its black color and
oily nature, diesel particulate may have a disproportionate
effect on soiling conpared to the effect of other types of
particulate (i.e., diesel particulate would produce more
soiling than TSP on a gram-for-gram basis). The black color
may make the soiling more apparent to the observer and the oily
-------
4-50
Table 4-11
Average Reduct'ion in Visibility
Due to Diesel Particulate
(percent reductions from base-year visibility)
1995 2000
City Size (population) Base Controlled Base Controlled
More than 1,000,000 18 12 22 12
500,000-1,000,000 7 5 9 5 .
250,000-500,000 5 " ~ 3 7 4
100,000-250,000 32 42
-------
4-51
nature r.ay rake it Tore difficult to clean. Th= net effect
;.'ould be to increase costs to the general public for rr.ore
frequent and more thorough cleaning events. However, because
of th^ paucity of scientific data on the physical soiling
effects of diesel particulate and TSP, no definitive statement
of these relationships can be made at this time.
There is a somewhat larger body of literature available
regarding the costs associated with various levels of soiling.
Summaries of this economic literature can be found in an EPA
report regarding the benefits associated with diesel
particulate control,[31] and in the Draft RIA. These reports
conclude that there are significant economic benefits to be
gained from control of diesel particulates with respect to
soiling.
4. Conclusions
Based on the above projections, EPA believes that diesel
particulate emissions are a serious environmental concern with
respect to their impact on various health and welfare aspects.
It seems apparent that significant reductions in heavy-duty
diesel emissions are an essential element in dealing with this
environmental problem. The stringent controls on heavy-duty
diesels and urban buses being promulgated in the final rule are
viewed as effective means of reducing the future growth in
particulate emissions.
-------
4-52
References
1. "National Emissions Report," National Emissions Data
Systerr of the Aeror?.etric and Emissions Reporting System, U.S.
EPA/OAV/M/OAQPS/NADB/MDAD.
2. "User's Guide to MOBILE2 (Mobile Source Emissions
Model)," U.S. EPA/OAR/OMS/ECTD/TEB, EPA-460/3-S1-006, 1981.
3. "Compilation of Air Pollution Emission Factors:
Highway Mobile Sources," U.S. EPA, EPA-460/3-81-005, March 1981.
4. "Rollback Modelling: Basic and Modified," Journal
of the Air Pollution Control Association. DeNevers, N. and J.
Morris, Vol. 25, No. 9, 1975.
- 5. "Diesel Particulate Study," U.S. EPA/OAR/OMS/ECTD/
SDSB, October 1983.
6. "A Comparative Potency Method for Cancer Risk
Assessment: Application to Diesel Particulate Emissions,"
Albert, R. E., E. Lewtas, S. Nesnow, T.W. Thorsland, and E.
Anderson, submitted to Risk Analysis, 1982.
7. "The Highway Fuel Consumption Model: Tenth
Quarterly Report," Energy and Environmental Analysis, Inc., for
U.S. Department of Energy, November 1983.
8. EPA Technical Report, "Motor Vehicle NOx
Inventories," Amy Brochu and Dale Rothman, EPA-AA-SDSB-85-03,
November 1984, draft.
9. EPA Technical Report, "Motor Vehicle NOx
Inventories," Amy Brochu and Dale Rothman, EPA-AA-SDSB-85-3,
March 1985, final.
10. EPA Technical Report, "Heavy-Duty Vehicle Emission
Conversion Factors, 1962-1997," Mahlon C. Smith, IV,
EPA-AA-SDSB-84-1, August 1984.
11. "1982 Highway Statistics," Federal Highway
Administration, U.S. Department of Transportation,
FHWA-HP-HS-82.
12. Argonne National Laboratory's ANL-83 Projections,
provided to Jim DeMocker, U.S. EPA, OAR, as part of initial
NAPAP review, January 1985.
13. "The Highway Fuel Consumption Model: Eighth
Quarterly Report," Energy and Environmental Analysis, Inc., for
U.S. Department of Energy, July 1982.
-------
4-53
14. "GM Challenges EFA Concern with Future
NOx/Particulate Emissions," J.E. Nolan and E.J. Neiderbu=hl,
General Motors Corporation, The Environmental Forum, November
1984. •
15. 1980 OBERS: BEA Regional Projections, Bureau of
Economic Analysis, U.S. Deoartment of Commerce, Washington, DC,
July 1931.
16. Letter to Mr. T.M. Fisher, Director, Automotive
Emission Control, General Motors, from Richard D. Wilson,
Director, Office of Mobile Sources, OAR, U.S. EPA, April 11,
1984.
17. Letter to Mr. T.M. Fisher, Director, Automotive
Emission Control, General Motors, from Richard D. Wilson,
Director, Office of Mobile Sources, OAR, U.S. EPA, September
25, 1984.
18. "Effect of Source Growth on Annual NO* Air Quality
in Urban Areas," T.Y. Chang, Ford Motor Company, APCA Journal,
Vol. 32, No. 5, May 1982.
19. EPA Memorandum, ' "Off-Highway NOx Inventory
Development: NEDS Methodology," Charles L. Gray, Jr., Emission
Control Technology Division, to Richard D. Wilson, Office of
Mobile Sources, March 5, 1985.
20. EPA Memorandum, "Stationary Area Source NOx
Inventory Development: NEDS Methodology," Dale S. Rothman,
Emission Control Technology Division, to Richard D. Wilson,
Office of Mobile Sources, March 1965.
21. Methodology to Conduct Air Quality Assessments of.
National Mobile Source Emission Control Strategies: Fina1
Report, EPA-450/4-80-026, Energy and Environmental Analysis,
Inc., Arlington, VA. (Prepared for U. S. EPA, Research
Triangle Park, NC), October 1980.
22. EPA Memorandum, "1981-83 SMSA Air Quality Data Base
for Nitrogen Dioxide," Richard G. Rhoads, Monitoring and Data
Analysis Division, to Charles L. Gray, Emission Control
Technology Division, January 11, 1985.
23. EPA Memorandum, "Comparison of Diesel Particulate
and NOx Inventories: MOBILES vs. NPRM," Amy Brochu, Standards
Development and Support Branch, to Charles L. Gray, Jr.,
Emission Control Technology Division, September 27, 1984.
-------
24. "Briefing Docur.ent" prepared for EPA Administrate r
William D. Ruckelshaus by EPA-'s Acid Deposition Task Force,
August 1, 1983, excerpted in the Environmental Reporter,
Septeroer 2, 1983, pp. 754-56.
25. "Review of the National Ambient Air Quality
Standards of Nitrogen Oxides: Assessment of Scientific and
Technical Information—OAQPS Staff Paper," EPA-450/5-82-002,
U.S. EPA, Research Triangle Park, NC.
26. "Regulatory Impact Analysis on the National Ambient
Air Quality Standards for Particulate Matter," U.S. EPA, OANR,
SASD, Research Triangle Park, February 21, 1984.
27. "User's Guide to MOBILE3 (Mobile Source Emissions
Model)," U.S. EPA/OAR/OMS/ECTD/TEB, EPA-460/3-84-02, June 1984.
28. Memorandum, "Summary Emission and Fuel Use Forecasts
for the Industrial Sector: Base Case for EPA Emission
Reduction Analyses," Craig D. Ebert, ICF, Inc., to Jeannie
Austin, OPA, U.S. EPA, Washington D.C., November 12, 1982.
29. "Controlling Airborne Particles," Committee on
Particulate Control Technology, National Academy of Sciences,
Washington, DC, 1980.
30. "Health Effects of Air Pollutants," U.S. EPA,
Washington, DC, June 1976.
31. "Health, Soiling, and Visibility Benefits of
Alternative Mobile Source Diesel Particulate Standards," Final
Report, EPA Contract No. 68-01-6596, Mathtech, Inc., Princeton,
NJ, December 1983.
-------
*' 1
APPENDIX
(Input Information: Tables A-l through A-10)
-------
Table A-l
U.S. U r ban VMT* (pillions cf riles/year)
Source
LDV
-Gas
-Diesel
LOT
-Gas
-Diesel
HDGV
HDDV
Buses
-Gas
-Diesel
19
604
169
35
16
3
82
.15
590
13
.82
166
3
.76
.55
.57
1
2
Years
1984
.65
.50
.23
.59
.22
.35
629
179
35
18
3
.25
614
14
.52
172
6
.80
.28
.76
1
2
.71
.54
.90
.62
.23
.53
1995
775.36
722
53
225.43
181
43
37.44
30.48
5.03
1
3
.19
.17
.73
.70
.37
.66
2
841
246
39
35
5
000
.77
759
82
.36
177
69
.64
.11
.75
1
4
.43
.34
.25
.11
.50
.25
Total
829.85
866.61
1073.74
1168.63
Based on MOBILE3 Fuel Consumption Model, January 21, 1985
-------
Taole A-2
iModel Year
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
1981
1980
1979
1978 ' -
1977
1976
1975
T O-» A
± y t "i - -
1973
1972
1971
1970
1969
1968
1967
1966
1965
Urban
2B
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.632
0.632
0.631
0.631
0.630
0.630
0.630
0.63Q
0.630
0.630
0.630
0.630
0.630
0.630
0.630
0.630
0.630
Fraction
3-5
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.633
0.632
0.632
0.631
0.631
0.630
0.590 '
0.550
n . «^sn
0.550
0.550
0.550
0.550
0.550
0.550
0.550
0.550
0.550
"of VMT
6
0.473
0.473
0.473
0.473
0.470
0.466
0.463
0.459
0.456
0.455
0.454
0.454
0.453
0.452
0.451
0.450
0.449
0.448
0.447
0.443
0.439
0.436
0.432
0.428
0.430
0.430
0,430
0.420
0.420
0.420
0.420
0.420
0.420
0.420
0.420
0.420
- HDDV*
7
0.396
0.396
0.396
0.396
0.395
0.394
0.394
0.393
0.392
0.391
0.390
0.389
• 0.388
0.387
0.387
0.386
0.386
0.385
0.385
0.383
0.382
0.380
0.379
0.377
0.370
0.350
0.340
0.340
0.330
0.330
0.330
0.330
0.340
0.340
0.340
0.350
0
0
0
0
8A
0.394
0.394
0.394
0.394
0.388
382
377
0.372
0.366
0.365
363
362
0.360
0.359
0.359
0.359
.0.359
0.359
0.359
0.359
0.359
0.358
0.358
0.358
0.300
0.241
0.241
0.241
.241
.241
0.241
0.250
0.259
0.268
0.268
0.268
0
0
83
0.176
0. 176
0. 176
0. 176
0. 176
0. 176
0.176
0. 176
0.176
0.176
0. 176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
0.176
Used in MOBILES Fuel Consumption Model to convert
nationwide VMT into urban VMT; based on MOBILES conversion
factor analysis.
-------
Table A-3
Urban Fraction jf VHT
Model Year
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
1981
1980
1979
1978
1977
1976
1975
1974
1973
1972
1971
1970
1969
1968
1967
1966
1965
2B
0.710
0.710
0.710
0.710
0.709
0.707
0.706
0.704
0.703
0.702
0.701
0.699
0.698
0.697
0.695
0.693
0.691
0.689
0.687
0.688
0.689
0.689
0.690
0.690
0.690
0.690
0.690
0.690
0.690
- 0.690
0.690
0.690
0.690
0.690
0.690
0.690
3-5
0.710
0.710
0.710
0.710
0.709
0.707
0.706
0.704
0.703
0.702
0.701
0.699
0.698
0.697
0.695
0.693
0.691
0.689
0.687
0.688
0.689
0.689
0.690
0.690
0.685
0.680
0.680
0.680
0.680
0.680
0.680
0.680
0.680
0.680
0.680
0.680
6
0.82S
0.829
0.829
0.829
0.819
0.809
0.799
0.789
0.779
0.772
0.765
0.757
0.750
0.743
0.732
0.721
0.709
0.698
0.687
-0.680
0.674
0.667
0.660
0.660
0.660
0.660
0.660
0.660
0.660
0.660
0.660
0.660
0.660
0.660
0.660
0.660
- KDGV*
7
0.775
0.775
0.775
0.775
0.767
0.759
0.751
0.743
0.735
0.733
0.730
0.728
0.725
0.723
0.715
0.706
0.698
0.689
0.681
0.670
0.660
0.650
0.640
0.630
0.630
0.630
0.630
0.630
0.630
0.630
0.630
0.630
0.630
0.630
0.630
0.630
.8
0.850
0.850
0.850
0.850
0.850
0.850
0.850
0.850
0.850
0.850
0.850
0.850
0.850
0.850
0.826
0.801
0.777
0.752
0.728
0.708
0.689
0.669
0.650
0.630
0.580
0.530
0.560
0.590
0.620
0.610
0.600
0.600
0.590
0.590
0.590
0.590
Used in MOBILES Fuel Consumption Model to convert
nationwide VMT into urban VMT; based on MOBILES conversion
factor analysis.
-------
Taole A-4
Liant-Dutv Diese-l Sales Fractions
"Best Estinace"
"Low Grcwth"*
Model .Year
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
•1988
1987
1986
1985
1984
1983
1982
1981
1980
1979
1978
1977
1976
• 1975
:1974
1973
1972
1971
1970
1969
1968
1967
1966
1965
LDV
.115
.115
.115
.115
.115
.115
.102
.089
.076
.063
.050
.046
.041
.037
.032
.028
.023
.019
.039
.060
.045
.026
.009
.003
.003
.003
.003
.002
.002
.001
.000
.000
.000
.000
.000
.000
LPT
.339
.339
.339
.339
.339
.339
.300
.263
.226
.188
.150
.130
.120
.110
.100
.090
.080
.077
.071
.056
.024
.013
.006
.001
.001
.001
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
LDV
050
050
050
050
050
,050
,050
050
,050
,050
,050
,046
,041
,037
,032
,028
.023
,019
,039
.060
.045
.026
.009
.003
.003
.003
.003
.002
.002
.001
.000
.000
.000
.000
.000
.000
LPT
.150
.150
.150
.150
.150
.150
.150
.150
.150
-.150
.150
.130
.120
.110
.100
.090
.080
.077
.071
.056
.0*24
.013
.006
.001
.001
.001
. 000
.000
.000
.000
.000
.000
.000
.000
.000
.000
"Low Growth" fractions used in diesel penetration
sensitivity analysis.
-------
Table A-5
Heavy-Duty Diesel Sales Fractions*
Heavy-Duty Truck Class
Model Vear
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
1981
1980
1979
1978
1977
1976
1975
1974
1973
1972
1971
1970
1969
1968
1967
1966
1965
2B
.300
.300
.300
.300
.300
.300
.300
.300
.300
.290
.280
.270
.260
.250
.232
.215
.197
.180
.162
.122
.081
.041
,000
.000
.000
.000
.000
.000
.000
.000
.001
.002
.002
.003
.003
.002
3-5
.300
.300
.300
.300
.300
.300
.300
.300
.300
.290
. .280
.270
.260
.250
.232
.215
.197
.180
.162
.122
.081
.041
.000
.000
.003
.005
.004
.004
.003
.003
.003
.012
.022
.031
.029
.026
6
.550
.550
.550
.550
.540
.530
.520
.510
.500
.486
.472
.458
.444
.430
.419
.409
.398
.388
.377
.309
.242
.174
.106
.100
..071
.041
.038
.034
.031
.054
.076
.082
.088
.094
.100
.105
7
.700
.700
.700
.700
.690
.680
.670
.660 •
.650
.640
.630
.620
.610
.600
.596
.592
.588
.584
.580
.589
.598
.606
.615
.578
.514
.449
.415
.382
.348
.341
.333
.348
.364
.379
.413
.447
3A
1.000
1.000
1.000
1.000
.988
.976
.965
.953
.941
.928
.915
.901
.888
.875
.878
.881
.883
- .886
.889
.865
.841
.818
.794
.770
.726
.634
.586
.540
.492
.482
.470
.492
.514
.535
.583
.632
SB
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
" 1.000
1.000
1.000
1.00-0
1.000
1.000
1.000
1.000
1.000
1.000
.960
- -.920
.920
.921
.921
.923
.925
.867
.809
.751
' .721
.690
Based on MOBILE3 conversion factor analysis.
-------
Table A-6
Fleet-Averase Heavy-Duty
Emission Conversion Factors ( BHP-hr/rrd ) *
Model. Year
1962
1963
1964
1965
1966
1967
1968
1969
1970
. 1971
1972
1973
1974
1975
1976
'1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
, --1994 —
1995
1996
1997
' Gasoline
1.29
1.31
1.32
1.33
1.35
1.36
1.37
1.37
1.37
1.37
1.37
1.34
1.31
1.28
1.20
1.12
1.08
1.05
1.01
0.98
0.95
0.95
0.95
0.96
0.97
0.97
0.97
0.96
0.96
0.96
0.95
0.94
0.94
0.93
0.92
0.92
Diesel
2.74
2.74
2.73
2.72
2.76
82
88
2
2
2.94
3.00
__3.08
3.15
3.19
3.23
3.27
3.23
3.19
3.07
2.95
2.84
2.72
2.'60
2.56
2.51
2.47
2.43
2.38
2.36
2.37
2.36
2.35
2.34
2.33
2.33
2.32
2.31
2.31
Based on MOBILES conversion factor analysis.
-------
Table A-7
Growth Rates and Assurr.ptions Used in FKM HOx Analysis
VMT Growth Rates*
(%/year, compound)
Vehicle 1982-1995 1982-2000
Class Urban Nationwide Urban
LDV*« +1.9
LOT** +2 .2
HDGV +0.4
HDDV + 4.7
Stationary
Source Category _
__ Stationary Point***
Off-Highway
Combustion
Stationary Area
• — • —
+1.9 +1.9
+2.2 +2.1
0.0 +0.6
+3.2 +4.2
Source Assumptions
Growth Rate
(percent/year)
+ 1.3
+2.5
+ 0.8
0.0
Nationwide
+ 1.9
+2.1
+ 0.2
+ 2.9
Discount
Factor
0.0
1.0
1.0
1.0
* *
Based on MOBILE3 Fuel Consumption Model, urban growth
rates also used in diesel particulate analysis.
Light-duty urban fractions of VMT are assumed to remain
constant with model year; therefore, urban and nationwide
growth rates are equal.
*** Stationary point source growth rate assumes a certain
level of future NOx control, based on ICF, Inc.,
projections; point source emissions included only in
nationwide NOx projections.
-------
T=ble A-8
SMSAs Modelled for NO." (Design Values, pprr NO,)
H?RM Analysis -
(1980)
Boston (0.050)
Chicago (0.060)
Cleveland (0.048)
Nashville (0.047)
Interim Analysis -,-
(1980-82)
Boston (0.036)
Chicago (0.052)
Nashville (0.053)
New York (0.036)
Philadelphia (0.046) Newark (0.045)
.Steubenville (0.040) Philadelphia (0.039)
•Denver (0.046) Seattle (0.048)
•Reno (0.048) Tucson (0.037)
Wash., D.C. (0.036)
•Denver (0.041)
•Reno (0.043)
FRM Analysis -
(1981-83)
Chicago (0.044)
Cincinnati (0.036)
Nashville (0.053)
New York (0.037)
Newark (0.040)
Philadelphia (0.040)
Pittsburgh (0.035)
Wash., D.C. (0.037)
•Denver (0.052)
•Reno (0.043)
Los Angeles (0.059)
Anaheim (0.045)
Riverside (0.042)
.£.
d
California SMSAs included only in the FRM analysis; future
projections based on CARS data.
NOZ concentrations at or above 0.040 ppm (75% of std.)
N02 concentrations at or above 0.035 ppm (66% of std.)
Interim analysis results presented in Motor Vehicle NOx
Inventories (Technical Report), and letters to T. M. Fisher
(GM) and Donald R. Buist (Ford), all contained in the Public
Docket.
High-altitude SMSAs.
-------
Table A-9
Low Altitude NOx Emssion Rates and Assumptions
Different Tnan MOBILES Values for
Emission Inventory and Air Quality Analysis
Base Case:
(2.3/10.7)
Controlled
Case:
(1.2/1.7;
6.0/5.0)
Vehicle Model
Emission Rate Useful
Type Year ZM[1] DR[2] SEA[3] Life[4]
LDGT 1987 +
LDDT 1978-80
1981-84
1985 +
HDGV 1987
1988
1989-90
1991-93
1994-96
1997+
HDDV 1987
1988
1989-90
1991-93
1994-96
1997+
LDGT
Cl. II a[6] 1987
1988 +
LDDT 1978-87
' Cl. I[5] 1988+
Cl. II a[6] 1988+
HDGV 1987
1988
1989-90
1991-93
1994-96
1997+ c
HDDV 1987
"1988
1989-90
1991-93
1994-96
1997 +
1.74
1.83
1.48
1.89
4.86
4.83
4.79
4.71
4.58
4.50
17.58
23.18
23.06
22.84
22.60
22.44
1.74
1.21
0.94
1.97
4.61
4.57
3.76
3.66
3.59
-
13.05
12.98
10.73
10.62
10.54
0.04
0.06
0.06
0.03
0.10
0.10
0.10
0.09
0.09
0.09
0.00
0.00
0.00
0.00
0.00
0.00
0.04
0.04
Same
0.03
0.03
Same
0.10
0.10
0.09
0.09
0.09
Same
0.05
0.05
0.05
0.05
0.05
40%
«» v
— —
40%
--
--
40%
40%
40%
40%
— —
— -
40%
40%
40%
40%
40%
40%
as Base
40%
40%
as Base
—
40%
40%
40%
40%
as Base
40%
40%
40%
40%
40%
Full
Half
Half
Full
Half
Full
Full
Full
Full
Full
Half
Full
Full
Full-
Full
Full
Full
Full
Case
Full
Full
Case
Full
Full
Full
Full
Full
Case
Full
- Full
"Full
Full
Full
[1] Zero-mile emissions (g/mi).
[2] Deterioration rate (g/mi-lOK mi).
[3] Selective Enforcement Audit.
[4] Certification to half or full useful life.
[5] Le'ss than 6,000 Ibs GVW.
[6] 6,001 - 8,500 Ibs GVW.
-------
Table A-10
High-Altitude tiOx Emission Races
and Assumptions Different than !:OBILE3 Values
for Emission Inventory and Air Quality Analysis
Base Case:
(2.3/10.7)
Vehicle
Type
LDGT
LDDT
HDGV
HDDV
Model
Year
Emission Rate Useful
ZMC11 DR[2] SEA[3) Li£e[4]
Same as Low Altitude
Same as Low Altitude
1987
1988
1989-90
1991-93
1994-96
1997 +
3.84
3.81
3.78
3.72
3.62
3.55
0.10
0.10
0.10
0.09
0.09
0.09
40%
—
40%
40%
40%
40%
Half
Full
Full
Full
Full
Full
Same as Low Altitude
Controlled
Case:
(1.2/1.7;
• 6.0/5.0)
LDGT
LDDT
HDGV
HDDV
Same as Low Altitude
Same as Low Altitude
1987
1988
1989-90
1991-93
1994-96
1997 +
Same as Base Case
3.65 0.10 40%
3.62 0.10 40%
2.97 0.09 -40%
2.89 0.09 40%
2.83 0.09 40%
Full
Full
Full
Full
Full
Same as Low Altitude
[1] Zero-mile emissions (g/mi).
[2] Deterioration rate (g/mi-lOK mi).
[3] Selective Enforcement Audit.
[4] Certification to half or full useful life.
-------
CHAPTER 5
COST EFFECTIVENESS
The cost effectiveness of an action is the measure of its
relative economic efficiency toward achieving a specified
goal. It is primarily useful in comparing alternative means of
achieving that goal. The cost effectiveness of the final
particulate and NOx standards analyzed in this report will be
the subject of this chapter. Before the final analysis, an
overview of the cost-effectiveness analysis in the Draft
Regulatory Impact Analysis (RIA) and a summary and analysis of
the comments received will be presented.
I. Overview of NPP-4 Analysis _
In the Draft RIA, EPA determined the cost effectiveness of
-the proposed standards in terms of the dollar cost per ton of
particulate or NOx emissions controlled. These values were
used to make comparisons with the cost effectiveness of other
mobile and non-mobile source control strategies.
To determine cost effectiveness, two pieces of information
were necessary: the costs and emissions reductions of the
strategies to be examined. The costs and emissions reductions
used were those associated with an average vehicle on a
per-vehicle basis, rather than the total costs and reductions
for the entire fleet.
The costs .used were those determined in the economic
analysis cf the proposed standards. The emission reductions
were calculated for each year of the vehicle's life by
multiplying the vehicle's miles travelled (V1T) by an average
per-mile emission reduction. The annual VMT values used were
those determined by Energy and Environmental Analysis adjusted
to reflect EPA's lifetime estimates. For heavy-duty diesel
vehicles, a composite V4T was calculated by sales weighting the
individual values for light, medium, and heavy heavy-duty
diesel vehicles (LHDDV, MHDDV, HHDDV). The average per-mile
emission reductions used were developed- using information from
the MOBILE2.5 emission factor model and the Diesel Particulate
Study.
Two approaches were used in calculating the
cost-effectiveness values for the -proposed standards: an
annual approach and a lifetime approach. With the annual
approach, costs were allocated: 1) to each year in which
emission reductions were produced, and 2) in proportion to the
size of these annual reductions. The result was a
cost-effectiveness value which is applicable at any point in
the life of the vehicle, as well as over the vehicle's entire
-------
5-2
lifetime. This approach allowed for comparisons on a
consistent basis with recent EPA cost-effectiveness estimates
for other mobile and stationary source particulate control, and
stationary source NOx control.
With the lifetime approach, the lifetime costs were
discounted to the year of vehicle purchase and then divided by
the undiscounted total lifetime emissions reductions. The
lifetime approach was only used in conjunction with the
proposed NOx standards to allow comparisons with past mobile
source cost-effectiveness studies, where only this method was
used.
Special considerations in the case of particulate matter
led to the determination of several different
cost-effectiveness values for each standard. Since the effects
of particulate matter are highly dependent upon particle size,
emissions reductions and cost-effectiveness values were
determined on a total, inhalable, and fine basis.* Also, since
the great majority of people who are exposed to NAAQS
violations for particulate matter live in urban areas,
emissions reductions and cost-effectiveness values were
determined on both an urban and a nationwide basis. For the
urban estimate the only change made was that emissions
reductions in non-urban areas were excluded? no changes were
made in the cost estimates.
II. Summary and Analysis of Comments . _ ._
There were very few comments received that dealt
specifically with the cost-effectiveness methodology and
procedures used in the Draft RIA. Comments received on the
cost-effectiveness estimates that primarily address either the
costs of control, or the emissions reductions obtained, have
been reviewed in the respective chapters on Economic and
Environmental Impact.
There remained only three comments specific to cost
effectiveness. The Department of Energy (DOE) presented its
own cost-effectiveness estimates that indicated that EPA' s
estimates may be somewhat low. In their methodology, the costs
used were the undiscounted costs of control, and the standards
considered differ slightly from those that EPA considered.
Total particulate is all suspended particulate matter
regardless of diameter, inhalable particulate is
considered to be all particulate matter less than 10
micrometers in diameter, and fine particulate is
considered to be all particulate matter less than 2.5
micrometers in diameter.
-------
5-3
Several environmental groups called attention to the fact that
the cost-3ffectiveness estinates for the 0.25 g/BHP-hr and the
0.1 g/BHP-hr standard for urban HDDEs in 1990 were equivalent.
Based upon this, they questioned EPA1 s choice of the more
lenient level of control. Finally, DOE took issue with the
alleged use of a 100 percent discount rate for NOx enissions
from elevated stationary sources. In their opinion, this
effectively renders any comparison of cost-effectiveness
between mobile and stationary sources meaningless.
DOE' s practice of using undiscounted costs appears
inappropriate to EPA, considering the basic economic concept of
the time value of money. Since DOE did not present its cost
estimates in a detailed fashion, it is not possible to
determine how much of the difference in cost-effectiveness
"values can be explained by this difference in accounting
methods. In any case, even if the basic technological
economic, and environmental concepts were the same, the use of
undiscounted costs will lead to higher cost-effectiveness
values. Since there is no apparent reason for using this
approach, it will not be considered further.
In the Draft RIA, EPA did estimate the same value for the
cost effectiveness of a 0.25 g/BHP-hr and a 0.1 g/BHP-hr
particulate standard for urban HDDEs in 1990. However, EPA did
indicate that it believed that in fact the more stringent 0.1
g/BHP-hr standard would actually be less cost effective, for
several reasons. The maximum benefit and least cost
applications would have already been used to meet the 0.25
g/BHP-hr standard (with averaging), so that subsequent use of
traps on additional engines might be somewhat less cost
effective. Other factors cited which argued for higher cost at
the 0.10 g/BHP-hr level are greater development costs, the need
to design to lower low mileage-target emission levels, the use
of higher quality components, the probable need for more
frequent trap regeneration, and the increased risks associated
with in-use compliance.
As will be seen in the updated analyses here and in the
Alternatives Chapter, the cost effectiveness of a 0.10 g/BHP-hr
standard does turn out to be somewhat worse than that of a 0.25
g/BHP-hr standard, confirming EPA1 s original position. It must
also be noted that cost effectiveness is only one factor used
by EPA in deciding between control options; technological
feasibility has been the primary basis for the decisions in
this rulemaking because of the statutory provisions governing
both the NOx and particulate standards. It was on the basis of
technological constraints that EPA decided against a 0.10
g/BHP-hr standard for 1990.
-------
5-4
Contrary to the assertion by DOE, EPA did not discount the
emissions- reductions from stationary sources of NOx anywhere in
its cost-effectiveness analysis. While some degree of
discounting emissions reductions based upon spatial
considerations may be appropriate in comparing the cost
effectiveness on an urban basis, where stationary sources have
relatively little impact upon breathing zone concentrations of
NO2» this would be less appropriate for regional scale
considerations. Therefore, this analysis has not discounted
the NOx emissions reductions from stationary sources when
making comparisons of cost effectiveness.
Ill. Updated Cost Effectiveness Analysis
A. Changes in Analysis
— There has been no basic change in the methodology used to
determine cost effectiveness. The bases for determining the
costs and emissions reductions to be used in the
cost-effectiveness analysis remain the same, with their values
changing only so much as the estimates have been improved.
In the Draft RIA, the differences between annualized and
lifetime cost effectiveness were explained thoroughly.
Mathematically, the difference lies solely in how the benefits,.
i.e., emissions reductions, are handled. Lifetime cost
effectiveness reflects the case in which the emissions
reductions are undiscounted; annualized cost effectiveness
reflects the case in which the emissions reductions are
discounted at the same rates as the costs.
Discounting emissions reductions assumes that the
emissions reductions are worth more at the present time than in
the future. For NOx, where exceedances of the ambient standard
for NO2 are projected in 1995 and 2000, but not presently,
emissions reductions may actually be worth more in the future
than they v/ould be now. In the case of particulate matter, for
which many areas of the country already exceed the ambient
standard, this is not the case, and it could be argued that the
sooner reductions are obtained the better. Thus, it is not
clear to EPA if, or how much emissions reductions should be
discounted over time. Therefore, the estimates of cost
effectiveness in the final analysis- are shown using several
different discount rates for the emissions reductions. The use
of various discount rates here allows for proper comparisons of
cost effectiveness to other mobile and stationary source
controls to be made, and for the sensitivity of the cost
effectiveness to the discount rates to be established.
In the Draft RIA, the costs and emissions reductions
estimates for the later year, 1990, standards were presented as
incremental to the values for the 1987 standards. This yielded
an incremental, or marginal cost effectiveness. In the final
-------
5-5
analysis of this chapter, this has also been done for the 1991
and 1994 • standards, and is referred to as the marginal cost
effectiveness.* In addition, cost-effectiveness values for the
combined standards have also been determined for these later
year standards, and are referred to as the total cost
effectiveness.
Updated estimates of the expected annual and lifetime per
average vehicle VMT are shown in Table 5-1. The VMT for HDGVs
has not changed since the Draft RIA, and the estimates for
line-haul truck VMT are the same as for HHDDVs in the Draft
RIA. The estimates shown for non-line-haul trucks for the
various model years were determined by taking weighted sums
across the VMTs for LHDDVs and MHDDVs as given in the Draft
RIA. The weightings were derived from projected sales
fractions in each year by class, and corresponding projected
diesel sales fractions.[1] As these change over time, so does
the average VMT for non-line-hauls as a whole. For LDT^ and
LDT2, separate estimates of VMT were derived, which was not
done in the Draft RIA. These were derived by taking the annual
average mileage accumulation rates in MOBILE3 and multiplying
each year's VMT by a survival fraction derived from the
registration data in MOBILES.[2] The VMT for urban buses has
been updated to reflect more recent EPA data.Cl]
The emission rates for the proposed standards vary over
the life of the vehicle and are summarized in Table 5-2. The
particulate equations were derived using the methodology as
described in the Diesel Particulate Study.[4] For NOx, the
values are derived from the MOBILES emission factor model, and
represent the actual in-use emissions including misfueling and
tampering.[2] The slight difference in form for particulate
and NOx reflects the differences in how the emission rates for
these two different pollutants are determined. Note here that
varying emission rates for each year of the vehicle's life are
being used in this analysis; in the earlier analysis an average
rate determined at the vehicle's half life was used. This
change leads to improvements in the accuracy of the estimates.
3. Results of Updated Analysis
The emission reductions and cost-effectiveness estimates
for the NOx and particulate standards are shown in Tables 5-3
and 5-4, along with the costs from the Economic Impact
chapter. These costs represent the net present value in the
year of sale to the consumer, using a 10 percent discount
rate. It includes both the first price increase and increased
1991 standards marginal from the 1988 standards, 1994
standard marginal from the 1991 standard.
-------
5-6
Table 5-1
Vehicle
Annual and Lifetime Per Average Vehicle \MT (miles)Cl]
HDDS Nbn-Line-HaulC2,3]
Age
1
2
- 3
4
5
6
7
8
9
"'10
11
12
13
14
15
16
17
18
19
20
LDT].
17,394
15,373
13.553
11,917
10,447
'9,127
7,944
6,884
5,937
"~" 4,986
4,239
3,574
2,983
2,459
1,996
1,587
1,226
910
633
479
LDT2
18,352
16,149
14,175
12,409
10,831
9,421
8,164
7,044
6,048
- 5,058
4,281
3,594
2,987
2,451
1,980
1,568
1,206
891
617
465
HDGE
15,590
14,040
12,630
11,000
9,960
8,210
7,060
6,050
5,170
4,339
3,570
2,960
2,410
1,960
1,680
1,250
- 980
750
520
340
1988
22,077
21,269
20,724
19,489
17,871
15,949
14,045
12,021
9,972
7,499
5,691
4,393
3,841
2,727
2,118
1,630
1,168
952
703
423
1991
21,971
21,163
20,618
19,388
17,777
15,864
13,969
11,957
9,721
7,460
5,662
4,371
3,818
2,715
2,108
1,623
1,163
949
699
421
1994
22,042
21,234
20,689
19,456
17,840
15,921
14,020
12,000
9,755
7,486
5,682
4,386
3,833
2,723
2,115
1,628
1,166
951
702
422
Bus
45,000
45,000
45,000
45,000
45,000
45,000
45,000
45,000
45,000
45,000
45,000
45,000
0
0
0
0
0
0
0
0
Line-Haul
64,720
63,790
62,850
54,870
47,700
41,000
35,310
39,320
25,910
19,510
17,840
14,910
12,130
9,870
7,980
5,310
4,970
3,790 '
2,890
2,070
Total 123,648
Ecpected
Lifetime
Miles
127,691 110,190 184,363 183,418 184,048 540,000 527,740
LIJ Urban fraction of travel for HDEVs:
Non-line hauls = .475, line hauls = .176, buses = 1.000, all = .288
£23 Changes in \MT by model year due to changes in relative total sales
fractions of heavy-duty classes IIB-V and VI-VIIA.
[3] HDDE all classes \MT can be approximated by taking a weighted average of
non-line hauls and line hauls. Relative weights are .635 and .365,
respectively.
-------
Base Case:
(no further
control)
1988 Standard
0.60 g/BHP-hr
1991 Standard
0.25 g/BHP-hr
GX C
0.10 g/BHP-hr
for urban bus
1994 standard
0.10 g/BHP-hr
5-7
Table 5-2
Annual Per Mile Emission Rates
(grans/mi le)
Particulate
Vehicle Type
Non-Line-Haul
Line-Haul
Urban Bus
Non-Line-Haul
Line-Haul
Urban Bus
Non-Line-Haul
Line-Haul
Urban Bus
Non-Line-Haul
Line-Haul
Urban Bus
Model
Year
1988
1991
1994
1988
1991
1994
1988
1991
1994
1988
1991
1994
1988
1991
1994
1988
1991
1994
1991
1994
1991
1994
1991
1994
1994
1994
1994
Emission
ZM[2]
1.1765
1.1233
1.502
2.1917
2.1784
2.1581
2.6586
2.6502
2.6334
1.0084
.9628
.9472
1.8786
1.8672
1.8499
2.2788
2.2716
2.2572
.4012
.3947
.7780
.7708
.9465
.9405
.1579
.3083
.3762
Rate[l]
DRC3]
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
. .0000
.0000
.0000
.0000
.0034
.0083
.0218
.0216
.0265
.0263
.0216
.0308
.0376
[1]Emission rates vary slightly with model year due to
changes in the conversion factors between g/BHP-hr and
g/mi.
[2] Zero-mile emissions (g/mi)
[3] Deterioration rate (g/mi x (years-0.5))
-------
5-3
Table 5-2, Conf d
Annual Per Mile Emission
(orams'/mi le)
Rates
Base Case:
2.3 g/mi LOT
10.7 g/BHP-hr HDE
1988 Standard
1.2/1.7 g/mi
LOTi, LOTz
6.0 g/BHP-hr HDE
1991 Standard
5.0 g/BHP-hr HDE
NOX
Vehicle Type
LDGT,
LDDT,
LDGTZ
LDDT 2
HDGE
HDDE
LDGT,
LDDT ,
LDGT:
LDDT j
HDGE
HDDE
HDGE
HDDE
Model
Year
1988
1988
1988
1988
1988
1991
1988
1991
1988
1888
1988
1988
1988
1991 .
1988
1991
1991
1991
Emission
ZM[2]
1.94
1.76
1.94
1.97
4.89
4.77
23.18
22.84
1.19
' 0.94
1.54
1.32
- 4.67
4.56
13.05
12.86
3.82
10.73
Rate[l]
DR[3]
.0136
.0030
.0136
.0030
.0132
.0122
• "• -.0000
-.0000
.0107
'.0030
.0107
.0030
.0132
.0122
.0050
.0050
.0122
.0050
[1] Emission rates vary slightly with model year due to
changes in the conversion factors between g/BHP-hr and
g/mi .
[2] Zero-mile emissions (g/mi)
[3] Deterioration rate (g/mi x 1000 mi)
-------
Table 5-3
Urban Particulate Cost Effectiveness
Option .
(g/BHP-hr)
1988: 6.60
1988: 0.60
1991: 0.25
0.10
for urban buses
Bus Only 1991
1988: 0.60
1991: 0.25
0.10
for urban buses
1994: 0.10
Discounted
Benefits (tons) Rate
Costs($) (1]
46
671-774
(625-728) (2)
1758(31
(1712)
966-1122(4]
(296-347)
0%
.026
.111
(.085)
1.217
(.991)
.137
(.027)
5%
,,021
.090
(.070)
.953
(.779)
.111
(.022)
10%
.017
.075
(.058)
.775
(.635)
.094
(.019)
Cost Effectiveness ($/ton)
Discount Rates for Benefits
0%
1770
6050-6970
(7350-8560)
1440
(1730)
7050-8190
(11000-12900)
5%
2190
7460-8600
(8930-10400)
1840
(2200)
8700-10100
(13500-15800)
10%
2710
8950-10300
(10800-12600)
2270
(2700)
10300-11900
(15600-18300)
[1J Costs represent net present value in year of sale of the total cost to consumer, using a
constant 10 percent discount rate.
(2) Figures in parentheses indicate marginal values from standard levels of 0.60 g/BHP-hr in 1988
and 0.25 g/BHP-hr in 1991 (0.10 g/BHP-hr for urban buses).
(3| Cost, benefits, and cost effectiveness for urban buses only.
14] This value calculated by taking a vreighted average of the heavy-duty bus and non-bus trap
equipped costs to go from no control to .0.10 g/BHP-hr in 1994.
-------
Table 5-4
NOx Cost Effectiveness
Option (g/BHP-hr) Costs($)(1)
LOT: 1.2/1.7, 1988
HDC: 6.0, 1988
HDG
HDD
HDE: 6.0, 1988
5.0, 1991
J1DG
HDD
28
36
7
69
79-166
(41-128)12)
21
(14)
137-311
(68-242)
Discounted
Benefits (tons) Rate
0%
5%
10%
'.108
1.494
.027
3.181
1.989
(.408)
.115
(.090)
.085
1.211
.022
2.578
1.610
(.328)
.070
1.021
.018
2.175
1.357
(.275)
.093 .079
(.073) (.061)
3.863 3.126 2.634
(.726) (.583) (.488)
Cost Effectiveness ($/t'on)
Discount Rates for Benefits
0%
263
24
278
22
40-83
(100-314)
183
(151)
35-81
(94-333)
5%
334
30
341
27
49-103
(125-390)
227
(186)
- 44-99
(117-415)
10%
405
35
417
32
58-122
(149-465)
267
(223)
52-118
(139-496)
[1] Costs represent net present value in year of sale of the total cost to consumer,
using a constant 10 percent discount rate.
(2) Figures in parenthesis indicate marginal values from standard level of 6.0 g/BIIP-hr
in 1988 for HDEs.
-------
5-11
operating coses. Where applicable, the total and marginal
values fc?r each standard level have been presented. For the
particulate standards, only the urban cost effectiveness is
given in this analysis as the nationwide value has not been
used in comparisons with other sources.
The costs and cost-effectiveness values presented here
represent the long-term values for each of the standards as the
fleet stabilizes in its response to the change in standards.
In the short term, the costs associated with the standards will
be somewhat higher as discussed in the Economic Impact
chapter. This would in turn result in higher
cost-effectiveness estimates in the short terra.
The discount rate used for the benefits can have a marked
effect on the benefits and cost-effectiveness values. As seen
in the tables, this results in a 40-60 percent increase in
cost-effectiveness values in comparing results using
undiscounted benefits and those discounted at 10 percent. The
cost estimates from Chapter 3 used a 10 percent discount rate.
Thus, the cost effectiveness estimates at a 0 percent discount
rate are equivalent to the lifetime cost-effectiveness values
as described in the Draft RIA, and those at a 10 percent
discount rate are equivalent to annualized cost-effectiveness
values.
C. Comparison to Other Control Strategies
1. Particulate
Table 5-5 presents an update of Table 6-5 in the Draft RIA
comparing the relative economic efficiencies of controlling
particulate emissions from other mobile and stationary
sources. Other than updating the values from 1983 to 1985
dollars, based upon the consumer's price index for new cars and
the producer's price index for industrial commodities (5.9
percent and 2.4 percent respectively) no changes have been made
in the estimates for other sources, taken from the DPS
report.[5] As in~the Draft RIA, the comparison is presented on
the basis of total, inhalable, and fine particulate; and the
stationary values have been adjusted to reflect the relative
breathing zone air quality impact of those emission compared to
that of diesel emissions.£3,43
The updated estimates of the cost-effectiveness values for
particulate control for HDDV are generally equivalant to those
in the Draft RIA. Therfore, as would be expected, the figures
in Table .5-5 suggest that HDDV controls remain quite favorable
when compared to stationary source controls, regardless of the
size of particulate examined. Only the control of wet cement
kilns appears to be significantly more cost effective than any
of the HDDV standards. Thus, it is a fair conclusion to state
-------
5-12
Table 5-5
Annual Cost Effectiveness Comparison
for Particulate Control Of Urban HDDVs and
•Other Mobile and Stationary Sources (S/ton)[1,2,3,4]
Sources [5]
Cement Kiln
Bus Only 1991 Standard
MODE 1988 Standard
LDDT (.26,1987)
HDDE 1991 Standard
HDDE 1994 Standard
LDDV (.2,1987)
Kraft Smelt Tank
Electric Arc Furnace
Borax Fusing Furnace
Industrial Boiler
Kraft Recovery Furnace
Lime Kiln (baghouse)
Electric Utility
Lime Kiln (ESP)
Part
Total
1
2270
2710
9530
9630
11100
11100
12300
9740
13600
29900
33200
48400
48600
77600
iculate Size
Inhalable
770
2270
2710
9530
9630
11100
11100
14700
15300
17900
42000
42400
59500
70000
96700
Basis [6]
Fine
1840
2270
2710
9530
9630
11100
11100
22200
15400
20200
126000
60500
93000
159000
156000
[1] Stationary sources are discounted to reflect their
relative ground level effect.
[2] 1985 dollars.
[3] Emissions reductions discounted 10 percent.
[4] For simplification, the midpoint of the ranges were used,
where applicable.
[5] Ranking based upon inhalable particulate values.
[6] See References 3 and 4 for other mobile and stationary
sources.
-------
5-13
that the HDDV particulate standards are quite cost effective
when compared to stationary source and other mobile source
controls.
2. NOx
Tables 5-6 and 5-7 present updates of Tables 6-8 and 6-9
in the Draft RIA comparing the relative economic efficiencies
of controlling NOx emissions from various mobile and stationary
sources. The values for the more stringent standard for LDVs
has been updated from 1984 to 1985 dollars by 2.4 percent,
based upon the consumer's price index for new cars, but are
unchanged otherwise.[5] The values associated with I/M
programs for LDVs represent more recent EPA estimates.[10] The
cost effectiveness values associated with the stationary source
controls of NOx have been updated to reflect more recent
analysis performed by EPA1 s Office of Air Quality Planning and
Standards, the South Coast Air Quality Management District, and
EPA Region IX. [6,7,8,9]
As with diesel particulate, the updated estimates for cost
effectiveness of the NOx standards are generally equivalent in
the updated analysis compared to the results in the Draft RIA.
The estimates for HDGEs have increased from $15 to $278/ton and
from $55 to $l5l/ton* for the early and later year
standards.[3] This reflects changes in the emission factors
from MOBILE2.5 to MOBILES and increases in the costs associated
with the HDGV NOx standards. Since the emissions reductions
and costs associated with the HDGE NOx standards are small f
even slight changes in their estimates can have large effects
on the cost effectiveness as has been seen to be the case.
The final NOx standards for LDTs and HDEs remain quite
favorable in cost-effectiveness comparisons to other mobile and
stationary source controls of NOx. The final NOx standards for
LDTs and HDEs have lower cost-effectiveness values than almost
all of the other mobile or stationary source control options.
If the stationary source NOx emissions were discounted to
reflect their relative ground level effect, as was done for
particulate, the cost effectiveness of the proposed LDT and HDE
NOx standard would compare even more favorably.
Using undiscounted benefits.
-------
5-L4
Table 5-6
Lifetime Effectiveness Comparison
of NOx Control for Mobile Sources
Cost Effectiveness
SourceU] ($/ton) [2,3]
HDDE 1988 Standard 22
HDDE 1991 Standard 35-81
HDGE 1991 Standard 183
LOT 1988 Standard 263
HDGE 1988 Standard . 278
LDVs (I/M, where presently exists for HC/CO) 527[4]
LDVs (I/M, where none presently for HC/CO) 2290[4]
LDVs (1.09 to 0.4 g/mi) - . . 2460[5]
[1] Ranked according to midpoint of range.
[2] -1985 dollars. -. - . .T._
[3] Emissions reductions undiscounted.
[4] See Reference 10.
[5] "Cost Effectiveness of Large Aircraft Engine Emission
Controls - Final Report," U.S. EPA, OAR, OMS, ECTD,
December 1979.
-------
5-15
Table 5-7
Annual Cost-Effectiveness Comparisons
for NO* Control of LDTs and HDEs and Stationary Sources
SourceCll
HDDE 1988 Standard
HDDE 1991 Standard
Industrial Residual Oil Boilers
HDGE 1991 Standard
LDT 1988 Standard
HDGE 1988 Standard
Industrial Coal Boilers
Internal Combustion Engines
Cement Kilns (Calif.)
Stationary Gas Turbine
Internal Combustion Engines (Calif.)
Glass Melting Furnaces (Calif.)
Refinery Heaters and Boilers (Calif.)
Cost Effectiveness
(S/ton)C2,3]
32
52-118
• 162C4]
267
405
417
456[4]
507[4]
812[5]
1010C4]
1320C5]
3550C5]
11200C5]
L1JRanked according to midpoint of range
[2] 1985 dollars
[3] Emissions reductions discounted 10%
[4] See References 6 and 7
[5] For applications in Southern California, see Reference 8
and 9
-------
5-16
D. Conclusion
The cost effectiveness of the particulate and NOx
standards is favorable when compared with other mobile source
control strategies. This is also true when these standards are
compared with stationary sources. Therefore, based on this
above analysis, the standards appear to be a cost-effective
means of reducing particulate and NOx emissions compared to
controlling these pollutants from other sources.
-------
' Appendix A
Summary and Analysis of Comments on the
Proposed Particulate Test Procedure for
Heavy-Duty Diesel Engines
-------
Surjtiary and Analysis of Comments on the
Proposed Particulate Test Procedure for
Heavy-Duty Diesel Engines
Following the publication of the NPRM, the HDD
manufacturers submitted written comments on the proposed
particulate test procedure. Also, a meeting was held between
the Engine Manufacturers Association (EMA) and EPA on January
28, 1985 during which HDD particulate test procedure details
were discussed. A memorandum describing this meeting is
available in Docket A-80-18. The written test procedure
comments as well as the verbal comments made at this meeting
are summarized and analyzed below in four groups.
The first group includes those which were well supported
by data or engineering analysis and which will not affect
measured particulate mass. The recommendation here is to
essentially accept the test procedure revisions contained in
these comments.
— - - The second group of issues include those which were not
well supported by available data or engineering analysis and
where the available data indicated that the change could
significantly affect measured particulate mass. The
recommendation here is to deny these requests for test
procedure changes, until it becomes clear that such changes
will not affect particulate measurements.
The third group of issues are those upon which -EPA
requested comment in the proposed rule, and the fourth group
are those which do not relate to Subpart N but are still
related to heavy-duty engine testing.
The analysis of each issue begins with a short description
of the aspect of the test procedure in question. The comments
made on this aspect are then summarized. Finally, the
available information relating to the issue is analyzed and a
recommendation is made.
I. Recommendations Accepted by EPA
Exhaust System Len-gth
Section 86.1327-87(f) of the proposed regulations
specifies that the distance from the manifold to the end of
chassis-type exhaust system should be a maximum of 12 feet.
Also, the length of exhaust system tubing from exit of the
chassis-type system or from the manifold to the dilution tunnel
shall be no more than 12 feet (maximum) , if uninsulated, or 20
feet (maximum), if insulated. This tubing shall be made of
stainless steel.
Summary of Comments: Ford is concerned that: 1) 12 feet
of chassis-type system may be too short for all in-use systems,
and 2) two maximum exhaust system lengths are possible,
-------
A-2
depending on whether a chassis type system is used (32 feet is
maximum) or if not (20 feet rraxirr.um) .
EMA expressed concerned about the following three issues:
1. "EPA has addressed the issue of exhaust system
design in the existing Final Rule for gaseous emissions (48 FR
52227) considering the effect of upcoming particulate control.
In Section 86.1327-84(f)(2)(i) of this final rule, EPA permits
a total of 32 feet length from engine to tunnel inlet."
"Engine manufacturers have all completed permanent test
cell installations following these guidelines. EPA has made
some significant changes in the current proposed rule (49FR <§
40314) that will cause significant modifications and undue
expense. EPA states that both a chassis-type and a
facility-type exhaust system may be used. It is not clear that
they infer "simultaneously." If EPA intends to permit only one
or the other system, then the individual lengths permitted
would require major test cell modifications to most facilities."
2. EMA is also concerned that the material that was
specified for the tubing is stainless steel which they believe
(a) is different from the gaseous emissions rule, and (b) is
not necessary.
3. EMA also requested that the rules exclude insulation
in the vicinity of instrumentation such as smokemeters.
Mack also expressed concern on the issue of exhaust tubing
lengths. Their position, while raised separately, is generally
.the same as the EMA position.
Analysis of Comments and Recommendation: The wording in
the proposed test procedure regarding allowable exhaust system
lengths is somewhat ambiguous. It was intended to specify a
total exhaust system length of 32 feet, with the option of
using either a chassis type system (with its own length
limitation), a facility type system or both together.
The final rule limits the amount of uninsulated tubing to
12 feet, which limits the amount of conductive cooling that can
be achieved from the tubing walls at a place where the
temperature differential is greatest. Yet, having up to 12
feet of uninsulated pipe provides reasonable flexibility for
engine changes without the incumberance of insulation. If the
typical length of an engines chassis exhaust is greater than 12
feet, use of the typical length is permitted, but only 12 feet
of it can be uninsulated.
A provision should also be made for up to 18 inches of
uninsulated tubing for instrumentation (an in line smoke meter,
-------
A-3
for example) since such instrumentation is required by E?A.
However, to maintain a consistent limit on uninsulated tubing,
such an uninsulated portion should be counted towards the
maximum total uninsulated length of 12 feet.
Based on EPA's experience, it appears that the type of
tubing steel should be irrelevant for diesel particulate
testing since it is soon covered with a layer of particulate
and further wall contact of the exhaust stream is unlikely.
The only exception would be steel with an extremely rough
surface which persisted despite a layec of deposited
particulate, which could occur if a rustable steel were used.
This could cause additional deposition. Thus, the tubing
specification should be changed to include typical in-use
-exhaust system materials, which could reduce costs for some
laboratories. However, the steel should be free from any rust.
Thus, in summary, it is recommended that the exhaust system
-specifications be changed and clarified to include provisions
for 1) a total length of 32 feet, 2) a system which can be
either chassis or facility type, 3) no more than 12 feet of
uninsulated tubing, 4) tubing in vicinity of instrumentation
can be uninsulated, and 5) tubing can be made of typical in-use
materials, but must be free of rust.
Dilution Air Filtering or Backpressure Measurement
Section 86.1310-87(b)(1)(iv)(B) of the proposed test
procedures requires the primary and secondary dilution air to
be filtered if background particulate is not measured.
Summary of Comments: EMA commented that if a manufacturer
does not filter dilution air or measure and correct for
background particulate, the manufacturer will only be
penalizing itself and not the environment (i.e., this will
cause a higher particulate emission calculation). Thus, EMA
recommended that the engine manufacturer should be given the
option to simply use good engineering judgment to account for
background particulate (i.e., filter dilution air, measure of
background particulate levels, or ensure backpressure levels
are sufficiently low so as to be ignored).
Ford also believed that need for filtering or background
correlation should be established by the manufacturer. . It
recommended monthly background checks, and if background
particulate is less than 1 percent of the standard, then it is
assumed to be zero and background samples need not be taken
with each exhaust sample.
Analysis of Comments and Recommendations: EPA believes
that filtering dilution air or accounting for background
-------
A--4
particulate levels is good engineering practice. However, if
background particulate levels are very lew, there will be a
negligible error in the emission results. In any event, any
error will only overstate true particulate emissions.
Therefore, it is recommended that the manufacturer be given the
option to control or account for background particulate as it
sees fit.
Calculation of Measured Particulate Mass
Section 86.1342-87 of the proposed regulations states "The
mass of particulates...is determined from the following
equation when a heat exchanger is used (i.e. no flow
compensation):
----- - ----- - p p ...
Pmass- (Vmix + Vsf) x ( _f - _bf ) x (1 - 1/DF)
V V
sf bf
Where:
Vmix - Total dilute exhaust volume (standard
conditions)
Vsf » Total volume of sample removed from the
primary tunnel
Pf - Mass of particulate on the sample filter
Pbf - Net weight of particulate on the background
particulate filter
; * '
Vbf » .. Corrected volume of primary dilution air
sampled by background particulate sampler
DF - Dilution factor
There are three issues here. They are: 1) should the
particulate mass on the filter plus the background be corrected
for dilution factor effects, or should just the background be
corrected, 2) should the calculation be based on Vmix or the
sum of Vmlic and V,f, and 3) which equations should be
specified for systems other than flow systems with a heat
exchanger.
Summary of Comments: EMA commented on all three of these
issues with the following statement. "The proposed equation is
both in-error and is inconsistent with all the equations
published in the Final Rule for Gaseous Emissions (48FR
p.52236) §86.1342-84(c) . In all the equations (1) through (4)
of this paragraph, HC, NOx, CO, and CO2 mass are calculated
based on Vmlx and not on the sum of Vm, x + V$f. V, f is
-------
A-5
not a significant portion of V.,,, x, typically V,f is less
than 0.1 percent of Vm, „ and can be ignored and it should te
just as it is in the final rule for gaseous emissions. Also,
all these gaseous " equations correct only the background
measurement by the dilution ratio effect. Therefore, the
equation for Ptnass should be:
Pmass
P P
_! bf
Vmix x V - V x (1 - 1/DF)
sf bf
Other sampling procedures will require different
equations, e.g., proportional mass flow control system and
systems where only secondary dilution air is filtered,
manufacturers should have the option to use alternate equations
compatible with their systems and good engineering practice."
Analysis of
Comments
and
Recommendation:
It is
technically correct that only background should be corrected
for dilution factor affects (this was a typographical error).
It is also true that the current equation only applies to
certain system designs. Thus, use of other equations that are
based on sound engineering principles, should be permitted for
alternate systems, but subject to prior approval with the
alternate system itself.
However, while V$ f is small for many systems, including
essentially all gaseous pollutant sampling systems, with some
double dilution particulate sampling systems it could be
significant. Therefore, Vsf should continue to be included
in the equation, if significant. However, little accuracy
would be lost if V, r were ignored if it was less than 0.5
percent of Vmi„.
Thus
the recommendation is that 1) sampling volume
(V,r) be retained in the equation, if it is less than 0.5
percent of Vraix, 2) only background be corrected for dilution
factor effects, and -3} other equations be permitted, if
approved in advance by the Administrator. -_
Balance Requirements
Section 86.1312-87(b) of the proposed regulation requires
that the balance used to determine the weights of all filters
shall have a precision and readability of one microgram.
Summary of Comments: EMA does not believe that the one
microgram balance is necessary because the accuracy gained does
not justify the additional expense and increased
associated with the one microgram balance. In
EMA's knowledge there are not any one microgram electronic
weighing time
addition, to
-------
A-6
balances available that have weighing chambers large enough for
the 90 TOT or 110 mm filters that are used on the EPA transient
test cycle.
EMA also presented the results of an analysis that was
conducted that compared the overall accuracies expected with 1
and 10 microgram balances. The 10 microgram balance was
analyzed assuming a precision of 20 micrograms. EMA concluded
that although the 1 microgram balance improves the filter
weighing accuracy by a factor 20, this accuracy is lost in the
particulate equation where other measurements are included that
have 1 percent, 2 percent, or even 3 percent uncertainty. The
net effect is that the 1 microgram balance, as compared to "the
10 microgram balance with a precision of 20 micrograms, reduces
the error by only .02 percent, from 5.30 percent to 5.28
percent. (This was calculated with a filter loading of 4 mg.)
EMA believes that this example illustrates the fact that there
is little benefit in having one measurement substantially more
accurate than other measurements used in the same process.
EMA also makes an argument about the cost of balances. A
typical 10 microgram balance costs approximately $3,000 but a 1
microgram balance costs approximately $7,000. They feel that
the additional expense of a 1 microgram balance should not be
forced upon manufacturers because the above analysis does not
justify it in terms of gained accuracy.
Analysis of Comments and Recommendation; EMA's analysis
of errors contained in their test procedure comments appears
"fundamentally sound. The affect of using a balance with a
pre'cision of 20 micrograms and a readability of 10 micrograms
appears minor and thus it is recommended chat the test
procedures be changed to reflect this.
Filter Reweighing
Section 86.1339-87 of the proposed regulations requires
. that if a filter is removed from the weighing chamber and not
used within one hour, it must be reweighed.
Summary of Comments: EMA sees no justification for this
requirement and recommends its deletion. They argue that
"there can be occurrences when an unscheduled test delay occurs
and filter and holder assemblies remain out of the weighing
chamber for more than one hour. During this delay, the filter
disc may be installed in the sealed holder and no changes in
dust or moisture content could occur. If the filter assembly
was installed in the test fixture during this delay and some
moisture penetration and deposition could occur, more moisture
deposition will occur during subsequent sampling of the exhaust
gas mixture. All moisture deposition either prior to or during
-------
A-7
sampling that is condensed on the filter will become adjusted
co the weighing room's moisture level during the stabilization
period prior to final weighing."
Analysis of (Torments and Recommendation: The purpose of
the rules regarding reweighing is to reduce water vapor and
particulate contamination of filters from sources other than
test-generated exhaust. If a filter is installed in a
completely sealed filter assembly, or a sealed filter holder
assembly is placed in the sampling line through which there is
no flow, then such contamination should be so negligible that
filters should be able to go up to 8 hours before they would
have to be reweighed. However, if these conditions of filter
placement are not met, then filters should be reweighed after 1
hour. Thus, it is recommended that the requirements be changed
to: 1) specify reweighing after 8 hours if the filter is in a
sealed holder assembly or in a sealed assembly mounted in a
sampling system through which there is no flow, and 2) specify
reweighing after one hour if the -above filter placement
criteria are not met. ...
"Sandwich" Filter Handling and Weighing
Section 86.1339-87 of the proposed rules requires that
both the primary and backup filter be weighed independently so
that the ratio of their net weights can be determined. The
backup filter net weight is deleted if it is less than 5
percent of the total.
Summary of Comments: EMA comments that "Some EMA members
weigh both primary and back-up filters together as a pair.
Then, after sampling, in removing filters from the holders, the
back-up filter is inverted on top of the primary filter placing
both faces with sample accumulation 'sandwiched* to the
inside. This procedure reduces the potential of lost sample
since now the filter 'sandwich* can be handled with tongs
anywhere including the center. This is especially desirable
with large diameter filters which tend to sag when supported at
the end. Weighing as a pair will, of course, reduce the number
of.required weighings, but will not permit the determination of
the ratio of the net weights which is the manufacturers
penalty."
Analysis of Comments and Recommendation; The procedure
that EMA discusses appears to be technically sound. Loss of
sample from filters that are weighted individually does not
appear to be a problem at the present, but the EMA procedure
appears to reduce the likelihood of sample loss even further.
The rule allowing a laboratory to not count up to 5
percent of total particulate filter loading due to particulate
-------
A-8
on the back-up filter is another point that EMA brought up that
also deserves analysis. whereas this has been a part cf the
HDD particulate testing procedures from their inception, it is
not good practice since it allows up to a 5 percent error which
could easily be avoided. This change will not be mace during
this rulemaking because prior notice has not been given and
some may consider it an increase in stringency. Nevertheless,
its elimination should be considered in the future.
The recommended action on this issue is that the
"sandwich" filter handling and weighing procedure be permitted.
Provision for Automatic Data Collection Systems
Section 86.1310-87(b)(5)(iii) of the proposed rules
specifies that "Chart deflections should be converted to
concentration before flow compensation and integration"
(underlining added).
Summary of Comments; Ford feels that this does not
account for automatic data collection (ADC) systems and
therefore, should be changed to include chart deflections an'd
analyzer voltage output.
Analysis of Comments and Recommendation: This section
dates from a period when ADC systems were generally not used.
ADC systems are now common and therefore Ford's recommendation
is quite reasonable. Therefore, it is recommended that use of
analyzer voltage output be permitted.
Hot-Start Restart for Reasons Other Than Engine Stall
The current regulations for gaseous emissions (Subpart N,
Section 86.1336-84(c)(3)) provides for a hot-start restart if
the engine stalls, but no provision is made for hot-start
restart after operator error or other small malfunctions that
can void a test.
Summary of Comments: Caterpillar suggested including test
voiding in the wording- for hot start restarts. They feel that
this would improve testing efficiency.
Analysis of Comments and Recommendation: The rules
regarding hot-start cycle restarts were revised to include
equipment malfunctions and were published in the Federal
Register as technical amendments on December 10, 1984. These
changes should adequately address Caterpillar's concerns on
this issue.
II. Recommendations Not Accepted by EPA
Six comments addressed aspects of the procedure which have
the potential to substantially affect measured particulate. In
-------
A-9
no case was there a substantial amount of data available upon
which to base a decision. However, in every case the available
data indicated that measured particulate mass could be affected
and thus that the current specification was necessary to
prevent biased measurements and unnecessary variability. In a
few cases, the analysis indicated that even the present
specifications may allow undue variability in particulate
measurements. These aspects of the procedure should be
reevaluated in the near future.
Location of Sample Line Temperature Specifications
Section 86.1310-87(b)(1)(i)(A) of the proposed regulation
specifies a maximum temperature of 125°F at the sampling zone
(in the primary tunnel) for single-dilution systems, but for
double dilution systems, the 125°F criteria applies at the
filter face.
_ ._ Summary of Comments: EMA recommends that the requirement
to be below 125°F at the sample zone for single dilution be
changed to refer to 125°F or less at the particulate filter, in
line with the double-dilution temperature requirement. The EMA
feels that this temperature limit is generic in nature and.not
dependent on the type of sampling system used; i.e., this
temperature limit and location should also apply to single
dilution systems.
EMA also presented data that they feel indicates that the
sample zone temperature has no influence on the single-dilution
particulate results. This data compares simultaneous samples
taken with a single-dilution system and a double-dilution
system. The single-dilution system had a peak sample zone
temperatures in the 220°F range yet peak filter temperatures of
about 110°F. The heat loss was. taking place in the sample
transfer tubing and filter holder. The average difference in
particulate mass results between the two systems was less than
0.5 percent.
In the EMA-EPA meeting of January 28, 1985, it became
apparent that the main issue here was the amount of heat
transfer that can be permitted in the sample transfer sections
of the single-dilution or, for that matter, the double-dilution
system.
Analysis of Comments and Recommendation: EPA's diesel
particulate sampling system specifications are based on several
precepts, two of which relate to the issue raised by EMA.
These are: 1) exhaust should be cooled to 125°F or less prior
to particulate sampling, and 2) this should be done to the
greatest extent possible by convection (i.e., using dilution
air) as this is the manner in which exhaust from an in-use
-------
A-10
engine is cooled in the atmosphere. The issue here is not the
125°F maximum temperature but rather how to achieve it.
The criteria for heavy-duty single dilution sampling
systems came from those for light-duty (LD) particulate
sampling systems, which is the area where most of the data
exist with respect to testing procedures. The light-duty
criteria (which is a single dilution system) is a maximum
temperature of 125°F or less in the dilution tunnel. This
reflects EPA's desire to maximize heat transfer by convection
(i.e., all cooling must take place in the tunnel) and limit
conductive heat transfer (i.e., heat loss in the sample line
cannot be used to reach the 125°F limit).
For heavy-duty (HD) particulate sampling, the same tunnel
maximum temperature of 125°F for a single dilution system
represents a direct extrapolation from LD experience and is,
technically, the most desirable system. However, for KD this
requires very large CVS systems (and large costs) and thus EPA
has allowed the alternate, double dilution system. EPA's
intent for this double dilution system is the same as for the
single dilution system; to achieve the majority of cooling
through convection. In establishing the specification .for
temperature (125°F) for this double dilution system, it was
applied to the filter face rather than the tunnel since all of
the tunnel flow is filtered and the end of the dilution tunnel
is essentially the same as the filter face (i.e., it does not
matter which is specified).
The data presented by EMA (see Table A-l) compares the
particulate results from a single dilution system experiencing
"a minimum of 110°F of conductive cooling to results from a
double dilution system which also appears to allow much
conductive cooling. The double dilution system used conforms
to EPA regulations, the specifications for which were made with
two purposes in mind. One was to limit conductive cooling.
The other was to allow reasonable lengths of transfer lines,
etc., for ease of assembly and location in the test cell. It
appears that the flexibility granted may have been excessive,
as. it was not the intent of EPA to permit excessive conductive
cooling from the double dilution system. Thus, at issue is not
so much the single dilution system specifications but rather
those of the double dilution system, which may have to be
modified in order to reduce the allowable amount of conductive
cooling.
While there is a limited amount of data which shows the
effect of conductive heat loss from sample transfer lines on
particulate concentrations (particulate increases as the degree
of conductive cooling increases),[1]* what is available
Numbers in brackets refer to References found at the end
of this section.
-------
A-ll
Table A-l
Simultaneous Particulate Sampling - EPA Transient Test
Cycle
Single
Dilution System
Peak "Tenp.'F
(Hot /Cold) Sample
(H/C)
Eiigine
1
2
3
4
5
6
7
8
9
10
11
12
13
14
.15
16
17
18
19
20
. 21
&igine
22
23
24
25
26
A -
C
- H
H
H
H
H
H
C
H
H
H
H
H
H
C
H
H
H
H
H
H
B -
H
H
H
H
H
Zone
Turbccharged
212
220
221
221
226
225
225
222
218
220
221
""' " 223
222
222
, „ :, 213
219
221
222
222
223
223
Turbocharged
299
302
214
174
153
Filter
6 Cyl.,
102
_ 109
109
110
110
110
110
106
111
113
109
111
110
112
103
109
111
111
113
114
114
8-Cyt.',
190
178
136
122
109
Part.
g/BHP-hr
4-Cycle,
.570
.604
.601
.591
.608
.613
.609
.602
.620
.603
.638
.651
.633
.625
.625
.613
.617
.609
.625
.624
.621
2-Cycle,
.389
.387
.427
.484
.464
Double-Dilution System
Temp,
Part.
q/BHP-hr
D.I. Diesel
.580
.603
' .591
.595
.620
.613
.606
.599
.624
.614
.635
' .629
.648
.633
.631
.625
.627
.618
.622
.637
.633
D.I. Diesel
.397
.381
.427
.460
.484
°F
Peak
Filter
104
107
109
109
109
109
109
100
103
105
101
103
102
104
94
94
103
104
96
104
106
-
79
78
79
78
75
Dilution
Air-In
80
83
86
85
84
84
84
76
- 80
82
77
78
77
78
73
76
79
78
80
82
82
89
90
92
92
87
% Dif*
-1.7
0.2
1.7
•-0.7
-1.9
0.0
0.5
0.5
-0.6
0.5
3,5
2.3
1.3
1.0
-1.9
-1.6
-1.5
0.5
-2.0
-2.0
" 1.6
O.n
5.2
-4.1
* Percent difference, single dilution compared to double dilution.
-------
A-12
indicates that conductive cooling should be limited to the
fullest extent possible. None of it argues for further
relaxations. Thus, it is recommended that no changes be made
in the specifications for single dilution systems, since this
system still represents that which is technically most
desirable. No tightening of the specifications for the double
dilution system should be made at this time, since none were
proposed. However, the specifications for the double dilution
system should be reevaluated in the future to determine if the
degree of conductive cooling currently allowed is acceptable.
Sample Flow Specifications and Proportionality
.Section 86.1310-87(b)(6), paragraphs (1)(B and c),
(ii)(E)(l and 2) and (ii)(G and H) , require that the gas stream
temperature into the particuiate sampling system flow
instrumentation and sample pumps be maintained at 77° + 9°F, and
also that certain temperatures be maintained within limits of
±5°F.
- The intent of these proposed requirements is to assure
accurate measurement of both the exhaust sample mass extracted
from the primary tunnel and the mass of the secondary dilution
air entering the particuiate system. This allows establishing
a means for maintaining the proportionality between the primary
tunnel mass flow and the extracted exhaust sample.
Summary of Comments: EMA and GM expressed in their
written comments that they believe that this section of the
regulations should provide system performance requirements, but
should not mandate the means by which such performance is
accomplished. -- -
In the EPA/EMA meeting subsequent to the submission of the
written comments, it became apparent that an additional major
issue of concern is the issue of proportionality between tunnel
and sample flow. EMA's position is that: 1) the proposed
regulations currently permit a ^5 percent deterioration in
sample flow from the set point for non-flow compensated
systems, and this same + 5 percent tolerance should be permitted
for flow compensated systems, and 2) a +2 percent flow change
specification is permitted for the main tunnel flow, and these
two flows (tunnel and sample lines) are independent and thus
the permissible limits should be added to permit a total of ±7
percent deviation from proportionality.
Analysis of Comments and Recommendation; The proposed
temperature requirements for particuiate sampling system flow
instrumentation and sample pumps are appropriate for some
systems but may not be appropriate for others. This can be
addressed by retaining the current proposals for sample flow
-------
A-13
handling and measurement but adding a provision that permits
alternate systems if these are shown to yield equivalent
results and if approved in advanced by the Administrator.
Section 86.1310-87(a)(7) contains a similar statement, but it
is not clear if it pertains to particulate sample flow handling
and instrumentation systems and, thus, the above clarification
will be useful.
The proposed rules are not adequately clear on the limits
of proportionality. The rules should be made explicit and
uniform for both types of systems (flow compensated and
non-flow compensated). The question is what should the
specifications be. •
EMA believes that tunnel flow and sample line flow are
independent and therefore the tunnel flow limits (±2 percent)
and nonproportional flow limits (+.5 percent) should simply be
added together to yield overall proportionality limits of ±7
percent. While the independence of these two errors can be
debated, the issue here is not equity, but accuracy. The
errors allowed for the currently specified system were derived
from the limits of equipment, not a decision that the errors
were the lowest desirable. Flow-compensating equipment
available commercially is capable of meeting a +5 percent error
specification at a reasonable cost. "~ The overall
proportionality limit of flow-compensated systems should,
therefore, remain at the ±5 percent level contained in the
proposed rules. However, this level of non-proportionality (±5
percent) may itself be excessive and should be studied
further. EMA has stated that they will be submitting data on
this issue, which should be useful for this purpose.
Thus, in summary the recommended resolution of this issue
is that 1) the proposed flow handling and measurement wording
be retained, 2) a provision be added that permits alternate
systems if these are shown to yield equivalent results, and 3)
a clarification be added which states that the ±5 percent
proportionality limit applies to both flow compensated and
non-flow compensated systems.
Test Cell Temperatures During Natural Cooldown
Section 86.1334-84 requires the test cell temperature
during natural cooldown to be 68 to 86°F.
• Summary of Comments: EMA states in their written
submission that:
"...None of the engine manufacturers have the capability
of cooling the test cells to assure that the natural
cooldown temperature limit can be met. If the limit can
-------
A-14
not be met, then a test may be postponed until the weather
cnanges. This practice is currently inefficient, but it
will become intoleraole when Selective Enforcement
Auditing becomes effective."
"In Section 86.1330-84 the cell ambient temperature during
the transient test is not required to be controlled for
engines which do not have temperature dependent auxiliary
emission control devices. The logic used for the cell
ambient temperature during the transient test should also
be applied to the natural cooldown."
In an EPA/EMA meeting subsequent to the submission" of
EMA's comments, this issue was discussed further. One aspect
of the discussion centered on the fact that two different
temperatures are specified in the Code of Federal Regulations
(CFR) at which a cold start emissions test can be started. If
the engine is force cooled, it cannot be started unless the oil
sump is at 75°F, yet if the engine is .naturally cooled it can
be started at 86°F. This requirement has been in place since
1984.
Analysis of Comments and Recommendation; The fundamental
purpose for cooling an engine by either natural or forced means
is to bring it to a temperature that is somewhat representative
of an in-use engines cold start. This is particularly
important for the measurement of HC and particulate emissions,
since emissions of these pollutants tend to decrease as cold
start temperatures increase. The temperature specification
that was selected for natural cool .down was 77°F, with a
tolerance range of +9°F. This was based on the current
light-duty practice. Ever, though the upper limit of this range
is 86°F, good engineering practice would dictate a target value
for natural cool down of 778F, and this is in fact the intent
of the rule. The fairly wide tolerance band is due to the fact
that most test cells due not have precise temperature control,
particularly in the summer.
A forced cool down procedure was added at manufacturers'
request to shorten the time necessary to prepare an engine for
a cold-start test. The upper temperature limit of 75°F for
forced cool down is consistent with the natural cool down
procedure for two reasons. Since it is relatively easy to
control the final temperature of a forced cool down, there is
no need to specify a wide tolerance band about the desired
target. There is no practical difference between 77°F and
75°F, particularly considering that the forced cool down occurs
much quicker than the natural cool down and, thus, some rebound
in temperature is likely to occur. Also, since forced cool
downs are performed to save time, it is reasonable to expect
that they will be stopped as soon as the required temperature
-------
A-15
is reached. Thus, if 86°F were the upper limit, this would
also be the average. The sane should not be true for natural
cool downs since manufacturers" are not expected to purposely
control the overnight temperatures of their test cells to the
upper-limit 86°F temperature. Thus, unless data are supplied
demonstrating that higher cold start temperatures have no
effect on emissions, it is recommended that no changes be made
in the cool down procedure.
Practically speaking, rejecting EMA's recommendation
should only have a minor economic impact on test costs. While
air conditioning test cells to ensure temperatures below 86°F
for natural cool downs can be quite expensive, this is not "the
only alternative available to manufacturers. The forced cool
down procedure can be used. Some manufacturers objected to
'this, due to the need to use city water to reach the 75°F
limit. However, internal cooling water can be used to provide
most of the necessary cooling and the cooler city water can be
used to provide the last 10-20°F of cooling. While
constituting some cost, the overall cost is less than that of
the water itself, since this water will be added to the cooling
water system within the lab and recycled.
Dilution Air Temperature Limits
Sections 86.1310-87
-------
A-16
"Due to the leadtine necessary to construct these
facilities, in anticipation of the gaseous FTP, the
manufacturers were led to believe that the CVS systems
constructed to meet these procedures would also suffice
for the impending particulate test procedures. To
redesign and modify these established systems in order to
add the necessary cooling capabilities would be a
difficult and expensive task for the manufacturers and
could possibly force the relocation of entire CVS
systems. An industry estimate ranging from $280,000 to
$420,000 has been obtained to equip test cells with the
necessary cooling capacity and controls."
The EMA is in support of the 125°F maximum temperature
requirement at the particulate filter holder. This
temperature limit effectively necessitates primary and
secondary dilution air temperatures to be significantly
below 125°F. In essence, the particulate filter
~~ "temperature requirement indirectly regulates the dilution
air temperatures to practical ranges. As suggested in SAE
Paper 800185,[2] little is known about the influence of
dilution air temperature on particulate measurements since
investigations to date have not separated the dilution,air
temperature factor from other dilution and sampling
effects. What can be said is that the combined effects of
, many of these factors on particulate measurements are
small, in the range of ambient to 125° F, suggesting that
any variations in dilution air temperature would have
insignificant effects on particulate measurements."
"The EMA recommends that the dilution air temperature
range (68 to 86°F) requirement be modified to allow
temperatures above 86°F, provided the dilution air is not
artificially heated above this temperature. This would
save the manufacturers the cost of adding cooling capacity
to their dilution air systems in order to provide for high
ambient temperatures occurring during warm summer months."
'-• • Analysis of Comments and Recommendations: EMA has
suggested that little is known regarding the influence of
dilution air temperatures on particulate mass concentrations.
While this is partially true, there are some data that show
that dilution air temperature is potentially a significant
factor. These data are presented by Reichel et al.,[l] where
they • show a 23 to 33 percent decrease in particulate
concentration when the dilution air temperature is increased
from 68°F to 122°F. When the dilution air temperature is
increased from 86°F to 122°F, the tunnel particulate
concentration decreases by about 17 percent. These reductions
in particulate concentration are primarily due to the
desorption of organics, according to the authors' theoretical
-------
A-17
calculations and thermogravimetric observations. EMA did not
specify how much in excess of-85°F they would like the upper
limit for dilution air temperatures and the dilution air
temperatures in the manufacturers' facilities would not likely
reach the 122°F of the above cited data. Nevertheless, the
data do indicate a significant effect on particulate
concentration due to dilution air temperatures.
EMA also refers to EPA's promulgation of the final rule
for heavy-duty gaseous emissions and they imply that this rule
also included all of the provisions needed for particulate
measurement. Actually, numerous changes in the final gaseous
test procedure requirements were made by EPA so that the
manufacturers would not have to invest in equipment needed for
particulate measurement at that time, if not so desired. EPA's
intent in so doing was to delay particulate testing equipment
requirements such that there could be a more ordered phase-in
for these equipment needs. One way of doing this was to
•attempt to assure that new equipment .that would be purchased
for compliance with the gaseous emissions testing rules would
also be useful when particulate testing was required. For
example, use of a dilution tunnel was allowed under the gaseous
emission regulations, but was not required. However, .the
gaseous rules and supporting documents did not imply that
particulate testing would not require additional equipment and
specifications such as secondary dilution tunnels, weighing
balances, and dilution air temperatures. An additional
observation on dilution air temperature limits is that these
limits have also been in effect for seven years of light-duty
particulate testing.
Therefore, since dilution air temperatures can have a
substantial effect on particulate emissions, no changes in the
proposed dilution air temperature requirements should be made
until some further date when sufficient data are available to
establish that no effect is present or to establish
satisfactory correction factors.
Humidity Effect Correction Factor for Particulate
Measurements
-x^
'No humidity-related correction factor currently exists for
particulate measurement.
Summary of Comments: EMA submitted a limited amount of
data on the effects of humidity on particulate measurements and
intends to submit additional engine data at a later date. The
current set of data show that the effect of humidity on
particulate is in the opposite direction and about
three-fourths the size of that for NOx. The equation would be
of the same general form as the NOx humidity correction factor
equation.
-------
A-18
EMA recommends that EPA consider a humidity effect
correction factor for particulate measurements using the
submitted data, with the option of accepting additional data at
a later date.
Analysis of Comments and Recommendation: The data that
are available on this subject are very limited (see Table A-2)
and are an inadequate base upon which to formulate a rule
change of the magnitude suggested in the EMA comment. In
particular, no data exist on the impact of various control
technologies (e.g., trap-oxidizers) on this effect. Therefore,
it is recommended that resolution of this issue await receipt
of additional data.
Sulfur Correction Factor
EMA suggests that a sulfur correction factor similar to
the NOx humidity correction factor be employed to correct for
the observed increase in particulate with an increase in fuel
sulfur.
Summary of Comments: EMA cites data that show that for
each 0.05 percent fuel sulfur mass increase, there is a
corresponding increase in measured particulate emissions of
0.024 g/BHP-hr due only to the change in fuel sulfur. EMA
suggests that a correction factor be used to correct for this
perceived inequity. Furthermore, EMA believes that a sulfur
correction factor will become more important as particulate
standards become more stringent in the future.
Caterpillar raised a similar concern about the inclusion
/if uat-or /uhinh i <; a 5<;nr*i al-«arl UJT hh «n1fat-
-------
A-19
Table A-2
Calculated "A" Values
For Particulate Humidity Correction Factor
Engine
Cummins #1
Cummins #2
Mack *1
Mack #2
Mack *3
Mack S4
Mack #5 __J._.
Mack *6
Mack #7
Mack #8
Mack #9
Mack 676
IHC #1
Cummins 903
DDAD 871
Caterpillar fl
Caterpillar 12
Caterpillar #3
Caterpillar £4
Caterpillar #5
Caterpillar *6
Calculated "A"*
+.0022
+.0022
+.00144
+.00099 ~
+.00042
+.00017
+.00255
+.00123
+.00275
+.00138
+.00107
+.00317
+.00107
+.00383
+.00303 . . ..
+.00204
+.00205
+.00094
+.00107
+.00113
+.00097
Avg. 0.00170
Equation for Correction Factor:
1
Mean Part.
(g/BHP-hr)
.66
.38
.56
.38
.45
.83
.50
.49 --•
.42
.35
.61
.64
.61
.79
.42
.67
.52
.40
.52
1.77
.56
Corrected Particulate = (
1 + A (Humidxty - 75)
) X Observed Pars, icu i ^ ce
-------
A-20
+0.05 weight percent sulfur or less. Taking Chevron's
relationship at face value, this change in sulfur levels could
result in a change in particulate emission levels of +0.024
g/BHP-hr, which is +4 percent of the 0.6 g/BHP-hr particulate
standard. While this effect would represent a greater percent
of a 0.25 g/BHP-hr particulate standard, use of particulate
control devices such as traps should reduce the size of the
fuel sulfur effect somewhat. Nevertheless, this degree of
potential variability is larger than generally desired.
The relatively wide specification for sulfur content
allows the sulfur content of the test fuel to change with that
of commercial fuel without requiring modifications to the CFR,
which are costly and time consuming. This flexibility is
intended, from EPA's point of view, and should be maintained.
Use of a sulfur correction factor would necessarily require
that a target fuel sulfur level be specified, essentially
removing this flexibility. As the sulfur content of EPA's
current (or projected future) test fuel is not markedly
different that that used to develop all of the particulate
emission data used in the technical feasibility analysis in
Chapter 2, retention of the current provisions does not affect
the feasibility of the standards being promulgated as long a-s
the sulfur levels of commercial fuels do not increase
dramatically in the future.
The issue of in-use sulfur levels is addressed in Chapter
2, as a number of manufacturers requested that in-use sulfur
levels be controlled to lower levels by EPA to allow use of
various aftertreatment technology. There it was determined
• that the feasibility of the final particulate standards was not
contingent upon this control. However, it was also indicated
that the control of commercial fuel sulfur content would be
.further investigated in the future as a means of controlling
particulate emissions. Investigation of the potential for
in-use sulfur levels increasing in the future is a natural part
of such a study. Thus, any potential for high in-use sulfur
levels, and thus, high certification fuel sulfur levels, to
cause the particulate standards to be infeasible will be
investigated at that time. In the meantime, with relatively
constant fuel sulfur levels, feasibility should not be an
issue. Thus, it is recommended that no changes be made to the
test procedures to account for the sulfur content of the test
fuel.
Caterpillar suggested the elimination of the inclusion of
water associated with sulfate in the measured particulate
mass. How this could be done is not clear at this point and
requires further study. However, ammoniation of the filtered
particulate is one possible approach. As discussed above, the
standards being promulgated are based on measurements which
-------
A-21
include such water. Removing\-the;CwaterRnow would either reduce
the stringency of the standards or require that the standards
be modified. Thus, no action should be taken with respect to
water measurement at this time. Further study may be merited,
however, if future sulfur levels increase* or if desirable
future control technology is found to affect sulfate, and thus,
water levels. This study should be coupled with the analysis
of future commercial fuel sulfur levels and their potential
control described above.
III. Issues Raised by EPA in NPRM
The NPRM requested comment on four issues because of
potential improvements were believed to exist in these areas.
These areas were: 1) the possibility of relaxing the cycle
performance statistics of horsepower standard error, 2) the
possibility of changing the primary torque measurement method
to an electronically compensated case load system, 3) the NOx
correction factor for humidity, focusing on the-adequacy of the
current factor for low NOx engines and 4) the addition of a
standard calibration procedure for HDGE throttle control
systems.
In general, EPA received little response to these issues.
From the comments that were received it can be concluded that
there is no dissatisfaction or known problems with the current
system. The only area that did result in the receipt of data
was the NOx correction factor, where the data presented
indicated that the current NOx correction factor was
.appropriate for low NOx engines as well as current engines (see
Tables A-3 and A-4) . Thus, as a result of the comments and
analysis of these four issues, no changes should be made in
these areas.
IV. Other Issues
The last group of issues do not directly relate to Subpart
N but will nevertheless be addressed here because they deal
with test procedures.
Smoke Standards ^--
EPA did not propose to eliminate the current smoke
standards when it proposed to add particulate standards.
Summary of Comments: Mack commented that an engine that
meets the 0.60 g/BHP-hr standard will easily pass the smoke
standards and therefore, the smoke standards are not needed.
They present no data to support this but state that it is based
on limited data.
-------
A-22
Table A-3
Calculated "A" Values For
NOx Humidity Correction Factor —
Engines With NOx Emissions Greater than 6.0 g/BHP-hr
Mean NOx
Engine (g/BHP-hr) Calculated "A"*
Previously Submitted Data**
Caterpillar fl 6.68 -.0017
Caterpillar 42 8.96 -.0025
Cummins 42 7.66 -.0023
Cummins *3 6.36 -.0017
Mack #1 7.85 -.0032
Mack #2 9.39 -+.0028
Mack #3 . -.0028
Mack *4 7.74 -.0037
Mack *5 • 7.44 -.0025
Mack #6 7.02 -.0024
DDAD #3 6.12 ' -.0029
Mack 676 7.47 -.0022
DDAD 871 7.66 -.0025
Additional Data
Caterpillar *i 9.11 _ -=0027
Caterpillar #2 7.98 -.0029
Average A = -.00259
* Equation for Correction Factor:
Corrected Particulate =(—-——7-5 r-r-. TT\)X Observed Particulars
L- +• A (Humidity — /a;
** Public Docket No. A-80-18, Parciculate Regulations for
H.D.D.E., "Statement of the E4A." September 13, 1982, Appendix
"D", p. 2, Table i.
-------
A-24
Analysis o£ Comments and Recommendation: Low particulate
emission standards may lcv;er srroke levels on average, but will
not necessarily guarantee smoke levels below the smoke
standard. This is because the two standards and their
associated test procedures are not mutually inclusive as to
intent and result. The purpose of the smoke standard is to
control worst-case smoke levels, whereas the purpose of the
particulate standards is to control average transient cycle
particulate. Since the engine operating conditions which
produce wo.rst-case smoke are not dominant in the • transient
cycle, a given engine could conceivably pass the particulate
standard and fail the smoke standard. The implementation of
traps may be the one particulate control approach that woruld
provide smoke control, since traps are effective under all
driving conditions. However, it is unlikely that all future
engines will be equipped with traps and the cost of running a
smoke test is quite small. Thus, it is recommended that the
smoke standards and their associated test requirements be
retained.
'OfficialTest Data
Paragraph 86.090-29(b)(3)(i) requires that the
Administrator's data shall comprise the official test data 'for
any engine tested.
Summa ry of Comments: Mack feels that there is a wide
variation in test results from facility to facility with no one
facility singled out as grossly superior or in error.
Accordingly, in cases where the manufacturer and Administrator
- differ by more than 10 percent, Mack recommends use of a third
laboratory as a referee.
Analysis of Comments and Recommendation: The designation
of EPA data as the official test data has been in effect since
the implementation of emission standard in the early 1970*s.
As no evidence was presented that demonstrates why the current
approach is inadequate, it is recommended that no change should
be made.
EPA Approved Equipment
Paragraph 86.090-29(b)(2) requires the manufacturer to
provide "...instrumentation and equipment specified by .the
Administrator...." (underlining added).
Summary of Comments: Mack commented that "In the past,
manufacturers have been allowed deviations from the
instrumentation specified in the Code of Federal Regulations
based on demonstrated equivalency. Mack feels that there is no
reason to abolish this practice and the flexibility that it
-------
A-25
allows. They feel that demonstrating equivalency guarantees
that the accuracy of the testing will not suffer. Mack feels
that the wording should be changed to "...instrumentation and
equipment approved by the Administrator..." to allow the
manufacturer the flexibility to install the instrumentation and
equipment in a manner most suitable to his operation.
Analysis of Comments and Recommendation; The requirement
for equipment specified by the EPA has been in place for many
years. This requirement provides EPA with the flexibility of
being able to specify use of a particular measurement procedure
or technique to enable a more confident assessment of the
emissions of an engine. Whereas this flexibility should be
retained, it should be pointed out that EPA has no intention of
being unreasonable in exercising this provision. To date, EPA
has rarely, if ever, exercised this authority to require use of
special equipment with respect to heavy-duty diesel testing.
Therefore, it is recommended that this provision be retained in
-its current form.
-------
A-26
References
1. "Influence on Particulates in Diluted Diesel Engine
Exhaust Gas," Stefon Reichel, Franz Pischinger, Gerhard
Lepperhoff, SAE Paper 831333, 1983.
2. "Experimental Measurements of the Independent
Effects of Dilution Ratio and Filter Temperature on Diesel
Exhaust Particulate Samples," J. S. MacDonald, S. L. Plee, J.
B. D'Arcy, and R. M. Schreck, SAE Paper 800185, Detroit,
February 1980.
U S GOVERNMENT PRINTING OFFICE 1985-554-048/10003
------- |